“Maybe some creative jobs will go away, but maybe they shouldn’t have been there in the first place,” she said, referencing a technology is that is inherently reliant on the works of those creatives to learn how to function on a very basic level.
@@MrTrueoriginality Maybe the Spielbergs and Picassos won't lose their jobs, but the next generation of those kinds of ppl will never arrive because no one will choose to go into those fields anymore. So one way or another, we'll get stuck with subpar AI junk
@@MrTrueoriginality ok this is probably bait but I'll bite. You currently have companies like Hasbro who owns Magic the Gathering pushing to use AI for their art. Magic in the past used actual painters and currently uses painters and digital artists to create the art for their cards. If this continues exactly the people who are talented creatives will be losing their jobs
the weather app i use has added an ai assistant and the suggested questions are like "is it a good day to take a walk" or "should i bring my umbrella" when the purpose of a weather app is to open it, see the numbers on the screen and decide how you're gonna go about the day depending on your own preferences about the weather
"Fight against AI as if you have something to lose because you do." As a translator, THANK YOU for this. Just saw an AI translation tool label itself as "the best translator in the world" and I wanted to vomit
As another translator, i feel the same 🫠. The amount of people i see reading machine translations of chinese novels, as a very specific example, and say "Yes this is good, comprehensible and enjoyable" is truly mindboggling (because it's false, they're terrible) and i just want to shove my notes from my entire bachelor's degree in their faces so they can learn what a good translation is 😭🙈
@@bleuumscarlett7977 While I agree that AI translations of, in my case, Korean light novels, are terrible, frequently that is the ONLY way someone who doesn't read the language can read more obscure titles, so I feel like there is some room in there. ..."The best translator in the world" is just blatant false advertising, though.
But the thing is, ai can't replace translators because there are certain contexts that can't be summed up and spit out. The only issue is this copy machine can trick you into thinking it's smart when it's just three existing sources in a trenchcoat. It can't create, or genuinely understand language, it just processes what exists and spits. For example, it may not translate certain threats because inappropriate content is usually filtered out. It can't accurately interpret subtle tone or unique phrasing.
หลายเดือนก่อน +30
The problem with translation is the "Good enough for me" attitude companies have. AI assisted translation companies priced me out of the corporate market. They don't care about nuance or substance. Here's a number of words, how much does it cost to translate them?
I'm not fully against AI translation because it is sometimes the only way for people to overcome language barriers, but also people need to be reminded of the value of an actual human translator and still employee that as often as possible. The AI won't understand the full context or the cultural expression that's done through language. An AI translator will never be the best translator in the world, what shameless marketing.
Instagram has started flagging legitimate photographer's work as AI too, which is so upsetting for creator's like me who spend so long perfecting our craft, and it further blurs the line between genuine photography/art and generative AI crap.
I can help!!!! You have to upload your image to "Imgur" or other image hosting service first before posting on Instagram to avoid getting tagged as AI. The reason is cause lazy instagram is simply tagging as AI any image made or editted in photoshop, how do they know you might ask? Well I'll tell ya alright, it's cause of the metadata of the image, you can see it if you open your image in a text program like notepad, literally the first line will be something along the lines of made in photoshop ver. xxxx etc etc followed by giberish
@@jorgemantilla609 I can see your comment. Remember: if you put a link in your comment, TH-cam will almost certainly delete it. IDK if that's what you did, but there are certain things that really trip the auto-moderation system.
run away from photography please. I just saw an AI app that makes bad photos turn into masterpieces. products, people, weddings, phone photos made look like the best photographer took them from different amazing angles. its been a nightmare for me too I wish things were different.
It is debatable to what extent photography is genuine to begin with. Basically every "good" photo is heavily edited and barely a representation of the real world. The same with video. How much of what you see on TV is "real"? Has it ever been? We thought it was. It felt like it... but AI generated stuff has made a lot of people remove themselves from entertainment and fiction; does it really add value to our lives? I think it no longer does. I would much rather watch a person talk about a subject on TH-cam than watch some fictional story or a pimped photos.
google adding ai results to searches genuinely terrifies me- esp bcuz nowadays, theres So many ppl who dont know better and will take it seriously. it's danger outweighs the funny parts for me. also i agree w the quote others have mentioned- i wish we dedicated ai to miserable/boring work instead of aiming to replace what makes us human
Google already prioritizes the top 12 or so websites they like (TH-cam, Reddit, Wikipedia, etc) so it's not like they haven't been sabatoging the already functioning search engine system they have made and kept years ago.
I saw someone say 'I want AI to do my laundry and dishes so I can spend more time making art' and I agree. Ive made art since I was 9 years old. Ive always wanted to create in new ways. Music, movies, books, painting, etc. Creative fields are honestly already so oversaturated right now that its almost impossible to live off of them as an average person. The idea of ai is terrifying for so many reasons honestly. Im sad for the future.
It shows a clear motive on their part. Ironically, automating hard physical labour is difficult, so they're just trying to nuke the creative work instead. It's evil. Plain and simple.
My youth group leader handed me a sketchbook one day because I wouldn't stop doodling in my journal. I've been making art ever since, even after I left the church. I have been an artist all my life, and was fully braced for art to be my hobby or side job at maximum. Now I'm worried about posting my own work only to try and compete against myself for the attention. Obviously, you should make art for the sake of making art, but I can't deny AI has been very disheartening to witness.
i'm so sick of how hard generative A.I. is being shoved down our throats. Its EVERYWHERE and is so fucking hard to avoid. As an artist it genuinely so upsetting.
As someone pursuing scientific/animal illustration, the influx of image generations of animals (birds especially) really pisses me off. Like they never look like an animal that has ever existed and always have comments full of people under the impression that they are real. I just think its so scary how easily it is already being used to spread misinformation. I know that when compared to other awful uses, inaccurate animals isn't the biggest deal, but its just so sad to watch people get misinformed and lied to about stuff relevant to my field of work.
@@KuncanDastner I sure hope so! I just hope it either gets regulated soon because of how bad its already gotten, or the bubble bursts and the AI craze dies out like NFTs. I know its wishful thinking but I'm crossing my fingers that generative AI will be widely considered "cringe" in a few years.
it just feels like another rich person interest, like """web3""" or the MeTaVeRsE. Its so silly and not at all believable that it could be a real part of our life. We haven't even been able to actually *solve* any economic issues with it like dynamic automation for jobs (particularly minimum wage ones, where the companies have the interests of the workers the least by far). Its so, so silly. I don't mean layoffs, of course but i think there should be something to be done. or maybe the socialism is getting to me
@@ornitharts If anyone needs something good to do with AI, I'm guessing we just need about 1 corn deepfake of Biden and Trump passionately and graphically doing it spread on Twitter and other social media like wildfire for both sides of the aisle to crack down on and regulate AI into the ground hard and fast.
Sam Altman is creepy. One of the reasons I think why he used Scarlett Johannson's voice despite her not wanting it is because consent does not matter to a lot of AI "tech bros."
Louis Rossman has taken the time to start calling all companies that seem to not understand consent RAPISTS. These companies that don't want to ask permission have a RAPIST mentality. 100% recommend following Louis Rossman
Sam Altman felt so entitled to a woman's voice that he took it from her even after her stating she didn't want him to. Basic consent, struggled with even in areas that are not necessarily explicit, are a huge red flag and demonstrates they wont be able to handle the worse stuff in an appropriate manner. He felt comfortable enough to do this to someone so in the public eye, imagine what he (as a member of the security team as well as the CEO) is going to be comfortable with letting others get away with, to people who have so much less support. Edit: spelling
@@vittripps so the point was to make the AI sound like an AI from a film, oh no! She doesn't own the concept of having a female voice, get over yourself.
@@technicolourmyles The person *explicitly* revealed that he was doing it based off of the movie "her" in which Scarlet's voice was used for an ai. Upon legal pressing to reveal how they made the voice they immediately got rid of it, implicitly revealing that they used Scarlet's voice in some way.
@@technicolourmylesyou know Sam isnt going to give you a reward for defending him online for doing shity things. unless you also feel some sort of kinship with doing shity things like him in that case I don't feel like you have much value to anybody other than ways to exploit them so I don't think you have much use in any conversation about this topic.
The amount of times Google’s “A.I.” assistant gave me an answer that was exactly highlighted right below it is infuriating. It’s redundant at best and outright deceitful at worst
You can dump google and try different search engines. Some of them still use the google jiggywhiz (I guess they are licensed, or a subsidiary of google already), but many don’t, have created their own systems. It’s a bit of hunting around, forums etc, but it’s good to have a small cache of decent search engines. There are plenty of people developing non-google search engines, and it’s worth finding and supporting them.
I have a theory that as AI continues to mess up images, and those images are uploaded, AI will simply continue to get more extreme and messed up as it references OTHER messed up images created by AI. It will become worse.
@@Delmworks Feel like maybe they could add meta data that is generated with the image that identifies it as AI generated and then have AI training software scan meta data for that field before adding it as training material. Though, it would also make them stand out from the crowd and be easier to filter out for other websites and people so I guess they wouldn't want that.
@Lewtable my first reply got deleted immediately, thanks youtube. Anyway, there's potential in using those tags in non ai images to protect them from data-scraping.
So much of these ai features added to websites/services that don’t need them just feel like bloatware. It’s like they don’t care about how the people who actually use their product feel and only care about investors and looking “impressive”
it's not just "like" that. It 100% *is* that. Most companies nowadays will literally only do something if it makes the shareholders happy. (Not arguing with what you said, just agreeing)
@@ragcat3732 that’s exactly it. That’s why a lot of fast food places don’t care about the customer experience anymore. Customers aren’t how they make money anymore 😬
and the time spent rigging in some cheesy chatgpt api calls is time taken away from fixing bugs and QoL issues that users ACTUALLY care about. ask anyone who uses adobe software or corporate game engines, they've been complaining about some form of this for years
Remember artists and others if you want to find an image without ai images clogging your feed try to use these phrases after your search. -"stable diffusion" -"ai" -"midjourney" -"open art" -"prompt hunt" and everyone PLEASE educate your parents, grandparents and older friends who may fall victim to scams or clickbait due to these images
That only filters out the content from AI users who are honest about it. Irregularstuff5290 has a good idea that could work. It's unfortunate that it ends up filtering out our ability to see new work. I feel a similar way about hiring new college graduates. If I find ChatGPT on a resume, it's going straight in the circular file, since it's probably AI generated anyway, but that just filters out the AI-generated resumes from people who are honest about it. Similarly, a resume from a college graduate from 2022 and earlier actually had to EARN their degree, and didn't have these tools to cheat. This will lead to age discrimination, through no fault of the applicant, since it will bias people to select 2022 and earlier graduates.
"if you dont want to see AI images, use these terms in your search that will make sure all you see is AI images" what the actual flip are you on about my guy
@@TheKayasto "minus this terms". For quite a while there was an option to direct your search results by making some commands in search bar, like site:sitename for searching on one site or -word to avoid results with this word. Not sure if this still stands since Google is total trashfire now.
It's "cheaper" for them to soak up water and eat up electricity than it is to pay people a living wage. Not like they consider the people they plan on replacing worthy of those two things anyways.
Anybody notice google search is literally broken now? I can't find a single thing anymore it's like it completely ignores what I ask for specifically and just spews AI garbage at the top further wasting my time
I've occasionally been trying to look up products, like a laptop with a specific processor, and it just doesn't care. None of the results are what I was looking for. If I type JUST the model number of a laptop processor, it still shows laptops, but with different processors. There's usually not even a way to tell until I click.
@malucart I've had this problem exactly. Like I used to be able to search the product with the right words. But now it takes hours or days to buy something because either it won't show me what I'm looking for or there's 100 listed at wildly different prices leaving me unable to decide at all..
Here’s a sentiment I saw someone express about AI that I totally agree with: I want AI cancer treatments. I want AI solutions to antibiotic resistant diseases. I want AI solutions to increase accessibility for people with disabilities. I don’t want soulless AI art to replace artists’ jobs with objectively terrible quality.
Well, it is really not about what you want. There are a lot of things I do not want. If people want it and are willing to pay the cost for a product, then companies will hop on that train and see how far they can take them. This will change the markets, but in time I am sure generative AI will not replace art. People will get bored of it and there is no true business model in it. All of these companies investing billions will not be able to create long term revenue with generative AI. It is a fad. Nothing more.
@@juliusfucik4011 I think they're pointing out the fact that companies DON'T go by customer demand, they follow what they think shareholders want to hear. Which is why we only develop disease cures and vaccines at a snail's pace. Remember how fast the covld vaccine came out? Every treatment could be developed that quickly if weirdo rich jerks found it profitable to fund.
Please consider that AI is fundamentally unusable for most healthcare applications. Due to the system-inherent "hallucinations" of answers, queries cannot be trusted now or in future. Furthermore, especially when deciding treatments, guidlines of different medical associations (like cardiology vs anaesthesia) will vary wildly and even sometimes contradict. The AI can't filter between "which association will have weird sponsorship ties" and will just spit out whatever it finds. (like the HAES controvercy a few years ago - when studies first showed a drug to be harmful, but newer studies negated this harm. Some of the studies were sponsored by the drugs manufacturer and some by the competitions manufacturer, it's a real clusterheck). The only really prudent thing is to read the journals for yourself. In diagnostics, AI tools could be helpful in very narrow fields. The problem here is that most diagnosticians rely on much more than the images of eg. a "CT-scan". Information like "What exactly ails the patient", "What body composition/implants/other healthcare issues do they have" are crucial. It's not just picture in - diagnosis out. I am hopeful for computing-supported solutions in research, but those again apply to narrow fields. AI cannot solve cancer, nor should it be the arbiter of treatment. I am sorry for the longwinded reply, this is just an important topic for me personally. I am happy to answer any questions on the feasability of AI in medical research, if you have any. Sorry again, I hope you have a pleasant day.
@@juliusfucik4011 I'm gonna be real, the current AI debate is missing the forest for a tree. Generative AI is not the biggest or most widespread issue with AI, and the scope of AI will far outscale just making images - and images alone is already millions of jobs and livlihoods. Don't get me wrong - my dad is a photographer and artist. I get that part. My issue is that everyone is so focused on the art debate that it lets it be dismissed by people outside the art space WAY easier. "Who cares what the whiny artists think, they need real jobs rather than drawing furry... 'media'" AI will not stop with generative AI. The approaches of making fun of AI for sucking also are terrible ways to approach it. Because eventually, as with most tech, *it won't suck anymore. And then your criticism doesn't exist.* AI is something that will hit tons of industries. Remember how many factory jobs got erased with the assembly line, then with robots? *Multiply that by god knows what.* All cashiers, customer service, call centers, attendants, truck drivers(which are millions alone and make up the backbone of our lifestyle), and so many more. Hell, "high skill" jobs are on the chopping block too. Tons of CEOs and project leads have already been talking about replacing their computer science staff, coders, developers, graphic designers, etc. - things people need work experience and degrees for - with AI as soon as it's viable. Writers, teachers, news organizations, etc. can all be replaced by AI. Watching this discourse makes me think of "First They Came" (Martin Niemöller) - and while you can definitely say that it's a morbid and hyperbolic comparison, the actual theme of the text makes sense when separated from who "they" are. With AI, you have to understand - the nature of modern companies is to exist solely to increase profit and cut costs. If you are not an investor, you are ultimately a cost. If you CAN be replaced, you WILL, eventually, be replaced. First they - AI, that is, are coming for bloggers, artists, drivers, cashiers, etc. Eventually, it will come for "you" - so SPEAK UP EARLY. This isn't just someone else's problem. Those that choose to interact with AI or feed it new data are only assisting it becoming competent faster. Nothing will ultimately stop this, because the government wants to be in the lead - but it can be kept away from our daily lives and restricted from the bad actors already using it to pump the internet full of junk and misinformation. by the way - we often hear about the "homeless crisis" - but with how many Americans live paycheck to paycheck, and how many Americans make up the jobs I listed - losing only one or two of those jobs, even ignoring the cascading effect on other entry level jobs - would instantly double or triple our homeless population/those without income. we're on a knife's edge. we can't all be cut costs.
I took an entry-level librarian civil service exam last fall. There were at least four questions on the exam about AI and how librarians could use it. We protested the inclusion of the questions bc as information professionals researching and organizing information is what we’re trained to do. What purpose do we have if we replace ourselves with AI?
Generative AI Apologists: “AI is just a tool for creatives, truly democratising artistic expression.” The artistic expression: “Buy my product in which I have ostensibly put negative effort.”
What I love is the same people who claim it's "just a more advanced tool for artists" will turn around and tell you that it's doing exactly the same thing as creative humans when it blends up all the data and generates a new image from it. Like, you can't have it both ways.
@@heli0ns Which is why I think they hate us (artists). They (think they) can't be us, because they have no skills, maybe they are jealous. But they also don't value art, so they aren't willing to invest in developing any skills.
@@Mors I remember once getting in an argument with a girl who said ai was good and the same as referencing. She was an artist, we were taking painting, the year prior we shared animation.
I ghostwrite and proofread for a living, and I dread the day that AI is competent enough at my job that it replaces me. I've already seen reduced work opportunities the past year or so
rest assured, generative algorithms need billions of words worth of content and theyve almost scraped the bottom of the barrel. its not gonna get much better than it is now, which still isnt great.
@@everfluctuating but its still good enough for slop (not in the sense of low quality work, just that it's conceived as a disposable entertainment product), and slop pays the bills for most professional creatives
@@ckorp666and most of the ones who get to do only the genuinely interesting work, the dream job stuff, only got there after years of working those 'slop' jobs first.
Me too. I'm a technical writer, and the job market is bleak. I even see a lot of positions that were 6-month contracts to train AI models. So temporary work to train tools to replace me.
What scares me so much about AI fans, is they are replacing natural and healthy forms of human expression with the click of a button. This will undoubtedly have a negative impact for several reasons, and they will most likely see more AI usage as the solution to the issues it creates
@@mattmaas5790 i think u may have misunderstood what i meant. I was talking about creating art and missing out on the slow dopamine release of expressing urself through non AI art creation, not about social media
@@mattmaas5790 It can absolutely get so much worse. Think of how bad it was, and consider: It is/was that bad because of how the algorythm was tuned to find what makes people mad and send them more of it, because it would get them interacting more, as a result forcing them down a pipeline of radicalization. Now, assume that the algorithm can read and register all of your posts, analyze your personality, and feed you that rage at ABSURD efficiency, or even generate more posts itself.
@Ryzard I genuinely think the AIs that we use for ourselves (think Her) (not meta websites built in ai), will be smart enough to help block out low effort and provably wrong information, and be able to serve you more interesting + personality enhancing content that today might be lost in the grain.
I legit feel bad for Neal Stephenson, because I know he's going to be asked dumb questions about whether he endorses whatever dumbass nonsense some techbro is claiming was inspired by his books for the rest of his life. Snow Crash and Diamond Age introduce a lot of cool-on-paper ideas and then make it really clear *why* they would suck in real life, and I honestly wonder how many of these techbros actually read them, and how many just read *about* them.
@@Djonin got curious from your reply; a search brought up Lamina1, describing itself as a metaverse for the “next generation of digital creatives”. Not only is Stephenson associated with it, he apparently _founded_ the whole deal. Not exactly crypto, but it still slots right into that Web3 ecosystem techbros just ditched for the AI train. There’s life imitating art, and then there’s this. Jeeeeeesus
@@sashakoshka He did a satirical comparison, he said it takes as much power to generate one of the images as water to fuel a farmer's field, or something to that extent.
In my mind it all comes back to the notion, that when we heard that industrialization was going to replace jobs, we hoped that it would let people choose the creative proffesions, and leave the simpler tasks to the machines. Not the complete opposite :(( And it happens so much nowadays, writers, editors, filmmakers, artists have their work stolen and are being let go, because AI is cheaper.
Im so glad AI companies fucked up their marketing because I was getting really scared that it wasn't a craze and would get integrated into our lives and destroy everything fundamental to making us human. But I really don't think it's going away so im scared for what's next.
@@solarydays What an extremely unnecessarily mean and ineffective way to explain reassuring information to someone who is worried about a widely worried about issue lmao. What a jerk. Next time just embrace the opportunity to assuage someone's worries out of normal human empathy if you think you know better. Seriously, weird tone.
@@solarydays they're not talking about lying, they're talking about saying the exact thing you said but not insulting people, it's really not that hard to understand
@@solarydays you are actually required to be nice to people who haven’t done anything to hurt anybody, that’s like a basic rule of socialisation and if you don’t know it maybe you’re the one that needs daycare
I got a CS degree, and part of it was a required ethics course. Good, right? Nope. The prof just wanted to talk about "The Fappening" because the title people gave that leak was funny, and he treated it like a joke. We think about ethics! Because they're funny! :,) Apparently.
They should require "how to verify claims about tech" classes for humanities and artist folks. About 4:00, this guy says how he thinks genAI works to generate pictures. It's completely baffling how he could be so wrong if he had put in even five minutes of study of the thing.
Video game companies claiming generative AI will replace humans is particularly pathetic when every gamer knows AI-controlled bots are worse than almost any human player. It's not a hypothetical, video games already have plenty of examples of software algorithms trying to emulate humans and doing a bad job at it.
Not totally related, but I think it's worth drawing a distinction between character AI in video games and the obscure generative algorithms (as in, the programmers didn't write that code which makes the algorithm make its decisions) "AI". Character AI is currently written as code by humans and it's an important part of the creativity and meaning in a game. It can be as simple as "see player, move towards player" and as complicated as a daily schedule and quest dialogue trees. It's puppetry, literally artificially making the pixels look intelligent. The strangest part of this new tech craze is that the people trying to sell it don't seem to understand that it's a puppet. A particularly shit puppet because it's not got any of the human creativity involved.
@@bitnewt oml, as an artist and programmer I had been personally insulted for making an animated background with css animations because another artist did not know the difference tvt I wish more people know how software works
AI now beats every human in Starcraft. The bottleneck is teaching a new AI the inputs, rules and boundaries without the support of the developer. Reading the screen and simulating a mouse and keyboard is the next frontier and a very active field with serious scientists.
Haven't watched the video all the way through yet so I'm not sure if this will get brought up later on BUT I feel like the problem is people view art as the end product, and nothing beyond that. Not saying that viewing art as the end result is inaccurate, but I would argue that it's an incomplete way of looking at it. Behind every painting, poem, work of prose, sketch, every short film, every song, every felt puppet and and spec script and clay sculpture and knitted sweater, heck, even every doodle some kid makes in the margins of their homework - all of those end products have some kind of thought process put into them. A lot of studying of the arts entails studying the thought processes of those creators - what techniques did this painter use? what about this cinematographer/director? what was the designer's purpose behind constructing these garments this way? - and how they use certain techniques/play with their respective mediums to evoke those feelings. AI art can't do any of that. It just haphazardly spits together a cobbled-together product based on a prompt, there's no soul or deeper thought behind it or anything, no matter how technologically advanced it could get.
Yeah exactly!!! It's the reason why don't like calling ai art "art". The fact person spend their own limited time to learn a skill or just made something for the sake of making something is what makes something art and technology doesn't have feelings or motivation s making their art valueless. The human part of art makes a child's stick drawing 1000x better than anything an ai could make
I find this very much so, as I know people that use ChatGPT in software to supplement their beginner skills in programming making them worse in painful ways
I’m in art school right now and every teacher, guest speaker, lecturer always says that the process is the most important part. All my classes demand to see the process of our work. Go onto any artist’s social media site and you’ll probably see shots of their process as they work towards a final product. And seeing other people’s process is a great way to learn and grow your own art. AI just can’t provide that process.
Hope to educate here - generative AI doesn’t go and search the internet when you type in a prompt. There’s a training phase, where the AI model is created by training it on an enormous amount of data. The training data has been scraped off the internet, which is how the model “learned” to generate troll posts. When you type in your prompt, you’re interacting with a static AI model, it doesn’t search any new information from the internet.
I don't give a shit when it is searching the internet. The fact that it is using the works of real people without their consent, and undermining the ability of those real people to make a living, makes it a leach on intelligence.
@@anzaia2164 The pre-processing step for the large models does "fact check" the images through a CLIP similarity score. Images that fall below that score are discarded from the training set, which also will discard the anti-ai art images from artstation. So the only way for the model to produce nonsense from troll posts, would be if the companies training the model are lazy. I can't speak to Google and OpenAI, but I know Stability did perform CLIP thresholding on LAION for their training subset.
@@anzaia2164 that would require manual effort though, which costs money - when the whole exercise (at least from the PoV of those funding this) is to produce something that *everyone else* will give them money for, not something that they have to pay loads of factcheckers to make good! (and, on a practical level, "what is true" is kinda messy and non-neutral - should "Palestine is a country" by considered true or not?)
The sooner this AI bollocks fades into the same obscurity the NFTs are already fading into, the better. The thing I'm worried about is what will the next get rich quick-scam be that these same people will push.
As someone AFAB, ai is genuinely so horrifying to me, the way so many men are so okay with using our bodies for their own pleasure and amusement and go so low to get what they want. Not to mention the whole phenomenon of people filming random women and them thoes vidoes ending up in the algorithms for deep fake prawn its so scary
@@Shephardsatan it's not that deep bc they dont perceive the humanity of the victims, all they care about is curve that make neuron go "woohoo". its literally dehumanizing
I'm an artist and I've completely stopped sharing my work online for the time being. I hate having to do it, but given Adobe, I don't even trust my work out there at this point. I really hope we get some sort of regulations soon. I've dealt with art theft like any artist online in the 00s/2010s, but this is getting ridiculous
Just a tip: in case you do decide to share your work online again and post on Tumblr, it basically sold all of its posts to an AI company and now everyone's posts are being used to train AI unless you opt out of it in your profile settings. Good luck. It's a scary world out there, but genuine human emotion and art will always be essential.
Try Cara! It doesn't allow any genAI images, and encourages using scramblers like Glaze. Glaze keeps the image the same, but "poisons" it for image scrapers
For the record, Sony retracted their "interview" with Neil Druckmann after he complained publicly that they misrepresented him and manipulated his words.
I want AI to do the "personal paperwork" types of things. Auto pay, taxes etc. oh and clean my room. Not take all the "fun jobs". Art, writing, design etc
So you're OK with accountants losing their jobs, as long as it doesn't affect graphic designers? Why is one group of peoples careers more important than anothers? People in other industries might enjoy their jobs as well.
@@SSValVik 1) Auto pay= pay my bills that happen every month? 2) Taxes technically a form that should be taught in highschool 3) personal implies small time, like a how to save for a new pair of shoes Oh I want humans to keep the more complex stuff. Like accounting for Government dept. That helps the economy, because there's paycheck for a human to spend.
@@SSValVik Accountants aren't losing their jobs because your average joe schmo isn't hiring them to pay their taxes. Accountants that aren't employed by a single big company are making their money doing accounting work for small businesses who don't have enough work for a full-time dedicated accountant. in Australia most people can pay their taxes easily without the need of an accountant because our tax agency already does most of the work for us and has programs that do calculations. there are still plenty of accountants with jobs,
@@flossimoth Goes for most of the modern world I think. The need to do your own taxes or hire an accountant to do it for you is one of the those things that is seriously archaic with the US, it's up there with still using cheques.
part of me wants AI to be ethical purely based on how cool the math is behind it, but the other part of me cant find a way which AI is useful for anything besides slop
@@Garf-the-true-and-evilhow can we know it hasn’t already seduced humans are the cancer, as is the plot with every great sci fi, and gives false diagnosis as a means of depopulation???
@@Atlas65 in this context it means that the work is derivative and bland. (maybe poor quality work too) It is something mass produced without much thought about quality
I recently started online school again and have been using google as I normally would to brainstorm and ask questions. The AI that has taken over google honestly sucks. It’s difficult to find actual, trustworthy answers even on Google scholar. 😖
Yeah, I agree. Every time I use any AI tool, I am ultimately disappointed. It’s just not that smart and not very good. Mostly hype. It’s ok at some pattern recognition but otherwise it’s quite dumb.
Ublock origin can block the ai answers, and I'm pretty sure it can remove some websites from your Google search (like quora or pinterest, so you don't get clogged with stupid quora bs or images with shitty quality and no source), it also removes ads, and honestly I feel like ublock is a necessity for browsing the web nowadays
Scholar is more for if you already know what you're looking to find. If you're looking for papers specifically (and already have a few), I would recommend Research Rabbit (it's entirely free) since once you input a source you can get a web of how it connects to your other sources and look at papers those sources reference/build upon. It also can integrate with Zotero, which makes chasing citations/assembling a bibliography easier.
took my words away! what’s so funny about technology is that it’s meant to make lives easier, right? so with AI, I have to ask what it’s benefiting? who gets to benefit from AI? Of course, it’s not the working class. I hope the AI craze goes away like anyone. I hope that it goes away like NGTs, and maybe it will! I’m so happy you’re back and that you’re in recovery, Duncan! I hope you’re feeling better
I have never seen or heard that AI answer to slenderman characters, so that dry delivery of “ilikemenderman” had me bursting out laughing in the office. Thank you for that.
I just want an AI assistant called “Computer” which I activate by saying “Computer, where is Lieutenant My Car Keys?” And it’ll give a sharp but matronly reply “Lieutenant My Car Keys is not aboard the My House”, it’d be great.
You can program that yourself right now if u really really wanted. Code an app that uses Computer as a wake word and queries an AI that either has visual sight or geolocation data of your lieutenants.
@matheussanthiago9685 yep. The ai would be there to interpret your ability to say *anything*, whereas you could simply harcode phrases and not use the ai endpoint.
I'm a member of book-related communities online, and the amount of AI pictures is disgusting. I even saw a woman selling NSFW, AI generated pictures of popular fantasy characters where it was clear the AI cut together real photos of unclothed people, but no real guarantee those photos were taken ethically
my thing with that kind of ai is that it can be used to generate images of children. and theyre simply scraping images from the internet, and that means *certain* kinds of illegal images have gotten into its database inevitably. its horrifying and disgusting
Amazon now has ai generated answers and i have to click on show me the actual real answers and reviews written by real people for my questions about the product. Very annoying.
Since we are dunking on AI, i can explain the pizza glue thing: The food you find on ads are often held together with non-edible materials. This is normal in the indistry, milk just doesn't look good under a camera so they put cereal pieces in white epoxy instead. A photographer for Pizza Hut might need to glue the ingredients together for a better shot. And here comes AI, who can't tell the difference between advice for a chef and advice for a photographer
ths biggest ally in the anti-AI fight is Sam Altman. launching chatgpt so long before it finished cooking and making AI a tech world gold rush before it's use cases had been proved will probably make the bubble pop way before it's sustainable. whoever is managing to use AI and wants it to keep being a thing should be screaming at Sam to quit.
The most important thing about being a good artist is having taste. You won’t get that from an AI. The images have a very distinct look to them because the program can’t make actual decisions.
I love how you make sure to go into detail about certain administrative features of a certain software to "tell us the full story". This is soooooo educational! TY!!!!! 💜😈
I don't think it's a suprise that someone who is peddling an AI that might in the future possibly do what they claim, but presently gets tens of thousands of people fired, just so the stock price will go up. It is no suprise at all that when someone says you can't use my voice. They just use it anyway, they do not respect other humans, all they care about is themselves.
One of the most infuriating things in my teacher education program (we are studying to be certified as classroom teachers) is the random pushes for integrating AI in our work. It makes me want to chew glass. There isn't anything that an AI can help do in lesson planning that couldn't be better done with the help of another human and some extra time, which is apparently anathema to the school system. Why are we banning students from using AI to write assignments meanwhile being encouraged to use ChatGPT to lesson plan???
Hey, I part-time teach ESL (English as a second language) and actually find ChatGPT to be useful sometimes when creating lesson materials. You can give it a source text to make shorter or easier, you can also use it for creating exercises focusing on certain grammar features while using specific vocab that you want the students to practice. Ofc you need to check it afterwards for nonsense, but it has its uses. I am also finishing my art teaching degree and my thesis is on using GenAI in school art classes. It's actually a really good debate starter and makes kids think and articulate their opinions (which for context is much less the focus of our education system than I feel it is in the USA - I'm from a formerly communist country and a lot of the "shut up and do the work" attitude is still quite prevalent with the older teachers so you want to encourage students forming their own thoughts as much as possible) and putting in the text prompt requires students to think about images and conceptualise the composition, mood, colours, etc. It is a great starting point for talking about kitch too. All to say that this technology is not totally black and white when you use it critically (ofc there are tons of ethical problems, not denying that.)
i lost my old job bc AI killed the industry and that really sucked. i just so happen to be a training cheesemonger rn lol so like things really turned around for me but those three months before i found out about the cheesemonger job were really depressing. AI pls let me keep my cheese
It’s so frustrating to me that this is now the public perception of AI (and totally warranted based on all that’s gone down!) I know lots of folks using AI in the science and medicine research space and it’s incredible. It truly is changing the pace at which we can research subjects that have been hard to access simply because the sheer number of data points is so vast it would take teams of people decades to sort through. This is super important when it comes to things like microbial mapping. The important thing to remember though is that this is a tool and experts in these fields do not blindly accept the patterns that the algorithms find. Many random coincidences will occur with data sets this large and it is the researchers job to determine whether the patterns spit out are actually rooted in science. The average person looking up recipes on Google is not trained to interpret AI results in any capacity. The use of AI talked about in this video is dangerous and I worry about the damage it causing leading to certain regulation decisions that may limit its use science by bioinformaticians due to the distrust people will inevitably have of it.
@@foxxxyg this is the part of AI i'm excited about, and it sucks that i can't discuss it without clarifying that i'm not talking about AI-generated photos of women with 5 boobs on facebook. like how did that become the average person's idea of what AI is
@@foxxxyg im sorry, but damage needs to be done to its reputation bc it earned that rep fair and square. Ai needs to slow down. This may be frustrating for its actual good use cases as you descrobed... but its doing enough damage to all of us in the meantime thats its flat out not worth it while its got big corpos trying to make a quick buck off it
It still impresses me how much creativity is devalued when it is needed all around us. Entertainment moves billions of dollars in several industries. Fashion, marketing, all design areas - just a few examples of creativity-dependent fields. And yet these MFing companies decided that the best way to invest their money was to pay artificial intelligence to steal every product of creativity and dump it all in one place without a shred of critical analysis on what's being stolen, or HOW for that matter. Art is not only important for self-expression, it opens our minds to think outside, think far away, think in someone else's shoes. But nobody wants to pay for that and I don't understand.
Another factor not mentioned in the video I think people will find very encouraging; several court cases have now ruled that AI “art” is not eligible for copyright!! While, of course, this would be unlikely to get rid of a lot of the most annoying parts of the AI slop getting pushed right now it *WOULD* mean that a lot of the jobs most currently at risk of would remain (at least as much as they are now) secure. This is also (AMAZINGLY) an issue where even if a single highest court of any country says “no, companies are god kings and can profit off of ~everything~” in our mass media-internationalized market all it will take is somewhere like the EU or *gasp* even China saying this stuff isn’t eligible for equal protection for media outlets of any size to functionally have their hands tied just to be able to *have* a viable product at all. Just this once, the courts really may be our saving grace.
This whole time they could have just used a pool of public domain content to train their models, or maybe they could have paid a bunch of artists who explicitly consented to have their work be used to train Gen-A.I. and we wouldn't be here. The fact that we're here by this point seems to show that, to a large population of people, most art and music is entirely worthless. That we can just look up a Picasso and suddenly have a Picasso to look at and study, the only cost of admission being the internet and bandwidth necessary to download a JPG. All the art forms that are immune to being freely experienced by anybody online, physical sculptures and live performances, are notably absent from every single dataset used to train Gen-A.I. and that's not coincidental.
If the language we used to refer to all of these pieces didn't casually treat them as synonymous, using the same titles for pictures of paintings and instead having the photographer publish their picture under the painter's name as an alias, I think that it might not have gotten so bad. Someone can just take a picture of Guernica and title it Guernica, then publish that picture under Pablo Picasso as an alias. Nobody knows it's an alias except for when they use context clues, it's never explicitly stated. A preconscious part of the brain often feels like it's experiencing the painting for completely free, that nothing had to be invested in order to witness all the painting had to offer. That philosophy of creation brought us here.
I think, more accurately, it's worthless to the kinds of people whose lives revolve around stock prices. If they can get away with something to save costs before the legal loophole is closed, that's not an ethical issue - that's an opportunity. If they can stuff the service into everything to pump its user numbers up when people don't actually go out of their way to use it all day, wow, now they can claim a big number to investors! From the start, calling it 'AI' makes the end game clear: Sell it, then jump ship before they realize that it's nothing to do with fictional superintelligences and certainly won't be making the 99% redundant. And if it does happen to do that, well, they've still won, they came out richer, aye? Also, y'know, that stuff with the voice actor. Yeah.
I have a bit of a conspiracy theory about that. Imagine someone trains a genAI on Picasso and uses it to generate, and sale, a fake Picasso. Given what people generally mingle in the art market, I'd say that this would lead to the trainer sleeping with the fishies.
19:58 I never really through about that until now but yeah, actually, it is kinda sus how every AI assistant is supposed to be read as female by default.
@@ayjay2287 it's because in scientific studies women's voices have been proven to be perceived as friendly, comforting , warm ECT. They are accepted in a way male voices are not.
@@susanlippy1009 I think it's also got a lot to do with just the dudes developing it wanting a woman's voice in their lives, or the general concept of a woman. It's not like it is inherently unique to AI assistants. A lot of boats have female names and most sailors are men. False equivalency perhaps, but I don't think it is unreasonable to say that men might just have an inherent trait that makes them more likely to name things after the opposite sex, especially if they are in an occupation where the lack of a interaction with the opposite sex is much higher.
It's also a feedback loop. If they're scraping everything on the internet, then they will be training ai on ai. It will only ever get worse as more ai exists in the internet.
they're already doing this. generative AI is already producing diminishing returns. it's very possible it has already reached its peak, that's why they're now resorting to even shadier tactics like stealing youtube videos and instagram posts.
There were also cases where certain people were trying to make an image generator off of the artwork of an artist who passed away😵💫 Or finishing an artpiece that was left unfinished because the artist's death....
Wasn't there also someone who used one of these to "see what The Unfinished Painting would have looked like if it was completed ", but it was literally complete as is, it was left "unfinished" and titled that way by Keith Haring on purpose before his death to make a point about the aids crisis, which was what led to his own health issues and death and the deaths of many others in his community. It was also just really weird for the person to not only used a generator to "complete" a purposefully incomplete painting, but also they could just look at literally any of his other larger works?? He made a lot of them? Anyway, point being, a lot of the people using these image generators to copy the styles of the living and deceased seem to really just have very little respect for the artists they claim to like the art of, and less then zero respect for the actual pont of their work
This video dropped just in time for me to listen to it as a bedtime podcast, you’re voice is so soothing dawg thank you for making these awesome videos 😭🫶
I don't know who you are, but I scrolled through the comments before watching this video, and your community seems so nice. So, I'll finish watching the video.
I was so excited to get a job in the video game industry as an artist, that's been a dream of mine since I was 9. I'm a junior in high school and NOW I'm gonna have to go find something else to do with my life. It's a shame, I really had things figured out :(
Don't fall into the mental trap that this will ruin your chance at your dream career. Gen ai has major copyright issues, studios have tested using gen ai in game dev pipelines to build complete games and say it offers nothing, Nintendo is the first of the big studios who just announced they wont be using gen AI. There are so many reasons to still keep working towards your dream. The only games that will be using this without minimum human intervention will be scam mobile games.
A lot of the push for Ai isn't about the AI output itself, but people investing in it wanting it to catch on so their stocks go up. It's like the guys still hanging onto NFTs, Gamestop, Doge coins, and Earth-2 land. It's gonna go to the moon guys, this is not financial advice.
There's a video by Knee Payne who used to work at LucasFilm, reviewing an AI generated animated video. This animated short was supposed to be original, but what's funny is you can clearly see Mike from Monsters Inc. in the background. The thought of AI screwing over Disney and other big corps through copyright theft is such delicious irony.
i think altmans actions with re to the 'sky' voice are at least partially explanable by, frankly, aversion to the concept of consent. the exercise of power over unwilling subject is its own reward
I'm an academic and did a workshop on the usage of AI in the classroom offered by my university. I raised the concern that in my field, all results generated by AI are genuinely backwards, representing the field as it was 30 years ago, and that this was markedly slowing our field down again in terms of innovative and critical work. Every time they said "oh isn't that interesting your concerns are totally valid anyway" and continue on showing people how to design lectures using AI. No thought behind those eyes, truly.
Lol, Google chrome was not the thing that used ai, it was google the search engine, that is the default for firefox, you have changed nothing from switching browsers
I've been using one of these things to help with my game and it's pretty much only good for that it's not good at creating things it's more like a whiteboard that will organize your ideas for you but you get out what you put in basically you still have to DO the work
@@BlueBeetle1939 i have mostly used gen ai to generate dumb images and to prove how messed up it is - one ‘anime’ image i generated literally stole Sayori’s face from DDLC unprompted- but i have genuinely tried to use it as a tool a couple times. As a student of civil engineering my lecturers are now peddling it as a way to find sources quicker, which it is sometimes good at, but there will sometimes be huge mistakes. I have to individually check each source to make sure it wasnt talking bullshit about the wrong thing because of course, it doesnt know what its saying. I’m checking sources anyway, but so much for ‘quicker’. As an artist i have also tried to use it for helping with inspiration (like 4 times most). Mostly it has no idea what i’m talking about and generates the most insanely bad things. However, the other day it finally showed me something useful. It captured what i was trying to draw in a way that i couldnt find elsewhere on the internet, for a the plans of minecraft build of all things. But the images themselves werent serviceable, loads of mistakes, inconsistencies etc. i tried my damndest to find the OG images it might have referenced but i failed. AI could help creatives more if it wasnt designed to replace us. And if it was able to show the steps of how it works, then we could find the original images and illustrators/artists/etc and get the actual human perspective that would benefit us. I have nothing to learn from AI, but the worsening of search engines and pushing of AI is wearing thin the ability to find those perspectives. Apologies for the essay, hope your game is going well!
AI generated outputs are hyped up now, but at the end of the day they're equivalent to bulk bargain bin items and their value should reflect that. They take no effort to produce and their training is all stolen data, not to mention it can't legally be copyrighted in any way as the outputs are not classified as human made. So why should the market value AI outputs above human made art, music, etc that _does_ have thought and effort behind them, just like any other hand made/human made item? The only reasons are because of the "new thing" intrigue, scams that hugely and _purposefully_ overvalue the price of AI, and companies that are not being held accountable by anti-trust or worker's rights trying to drive their profit line up for their shareholders, thus pushing to have their human workers replaced by cheaper means without considering they could be making what is known as the "Schlitz mistake" and devaluing their brand with lower and lower quality until the point of collapse. Despite what some people who buy into the hype might say, current AI cannot truly replace anyone. It sucks at coding, it hallucinates, it makes basic artistic mistakes that no human would ever make. Doesn't matter if it's homemade software or a big company's AI on a stupidly overpriced subscription model; AI will always need a human supervisor to correct its mistakes, which defeats the point of having the AI do the job to begin with when you _already have a human available that can do it's job!_ The outputs are always guaranteed to be median - just totally average, because that's what the software actually does to the training data it's provided with; it finds common code between every image, written word or musical riff and makes a completely homogenised and predictable output every time. And that's not even considering _AI incest,_ where indiscriminate scrapes of data can cause the AI to feed off itself and become so inbred it's useless. It's boring and it's cheap, and when more people get past the hype, just like they did with NFTs and every new piece of "revolutionary" technological garbage before that, its value will drop significantly, leaving human made art still on top. That's not to say it can't do some damage in the meantime, or that it won't cause further damage in the future, but never, _ever_ buy into the idea that your work and your art isn't more valuable than AI. After all, the AI is trying and often failing to copy _you,_ not the other way around!
We have been forced to brave the seas in search of new booty ever since she sank. Can't even edit a pdf without paying an arm and a leg. CSP also unfortunately adopted the subscription model. No more one time payments.
AI is such a double edged sword. For medical applications and programming, it can be brilliant. For gamifying capitalism it will destroy our country from within.
I’m hoping that, like nfts, bc there hasnt been an immediate/rapid increase in profits for ai companies, that investors are going to pull out bc line isn’t going up fast enough. It’s arguable that line won’t go up fast enough any time in the near future because of how energy intensive ai is.
This is the first video I’ve seen you post in a while. Just an fyi like a TH-cam issue not you problem cause I watch all your videos. LOL time to binge I guess. Gotta catchup
I get closer and closer to "smash phone" every day. I work(ed) in tech and am seriously unsure I wanna go back into that industry. It's just kinda doomed and not always for obvious reasons. The one that really gets me is how DUMB those in charge are. It's cloud & blockchain all over but infinitely dumber cause people genuinely don't understand the tech but want digital divinity. It just... won't. Tech is hard. Real hard. None of this works how those in charge think it does. My gods it's such a slow moving disaster.
The only generative ai I “enjoy” is the weird shitposts that come from someone who is generally critical of ai. I find the slop it produces to be funny (somewhat), but I still don’t support ai (Great example of what I mean is the minions at the end, and it’s a “you’re only slightly more tolerable than other generative ai because you’re slightly funny”) This better make sense to other, I wrote this at 1 in the morning with a raging headache when I should be sleeping
I think the only way AI-generated images work as art is if there's an awareness of the fact that AI gives an uncanny, off-putting vibe and isn't a replacement for human work. The throwaway shitposts can be funny and I've also seen some interesting AI surrealism that leans into the "created by a creepy robot" feeling. You have to understand that you're dealing with something that's soulless, basically.
The one good thing that has come from all this AI generated slop flooding the internet is that I appreciate art more than I ever have. It makes me so happy to see creative works made by people who put in a lot of effort, the idea of every little detail being made by the hand of a passionate artist, with intention and a goal to put their imagination into the real world, it's so moving to me. I was at a convention recently where they had booths for artists to sell their prints, stickers, keychains, plushies, handmade crafts. It fills me with such warmth to see someone's face light up when you say you like their art, knowing that they have spent so much time perfecting their craft and probably worrying sick about whether it's good enough, and yet it always is. I feel a huge sense of pride being part of this community, and so I will proudly give my money to these artists, knowing it will fund their work and help them improve and flourish. Art is such a beautiful thing, and I believe it will prevail in the end. AI images, voices and text have no place in the creative world. Despite all its advancements it will always fundamentally lack the most important thing; the artist.
Remember when AI bros drop the old "AI's here to stay," that doesn't make it guaranteed to succeed or thrive. The way it's going it's looking to be the mark of lazy/scummy/talentless grifters, frauds, and amateurs.
Dear algorithm, all the commenters have already said pretty much anything I could add to the discussion here, but I think this is an important video. People keep telling me that commenting boosts engagement which will help videos do well in the algorithm, so please take this comment as my genuine plea to push this video to many other people. Thank you,
Censored "g*me" got me rolling. Truly, genAI is my personal Devil and I am working harder than ever to avoid it. BUT YOU CAN'T. Corporate greed is even WORSE now that these devices are part of our lives in and out of work.
It's not surprising to me that AI is already losing steam. My immediate reaction to seeing AI, like the very first time I heard about it, was "didn’t we learn nothing from NFTs like a year ago?" I've always seen this as being equally stupid as NFTs and blowing up because of tech bros DESPERATE to be the next innovator or whatever and falling down catastrophically once the tech companies finally realize nobody else wants this but them and The Bros™️. I can't wait for it to fall off. The fact that it's called AI in the first place has been driving me crazy for so long because it's literally not artificial intelligence. It got off on such a bad foot, it has literally 0 upsides or positives, and I hope it dies like it deserves to. On a somewhat unrelated note, I don’t know if it's a premium only feature right now or if it's something everyone can see, but TH-cam live streams have an AI generated summary of chat that they're experimenting with. It's kinda hilarious how bad it is. I watched a stream of people playing Mario Party the other day and I lost track of how many different games the AI summary thought they were playing (to be fair, one of them was Mario Party. But it also thought Among Us, Rocket League, League of Legends, Mario Kart, Super Mario Bros, the list goes on). It was really funny when you would see it blatantly copy messages that were completely unrelated to the stream but take them as fact
I’m surprised they don’t pick and choose what to train the AI with? Like if chicken is what we want them to learn from, why not pick and choose the chickens to show it? Wouldn’t that prevent the weird sarcastic pictures we have?
This video is incorrect on a number of technical points about how the AI image generators work. Very early image generators may have just used scraped datasets, mostly as a proof of concept that this sort of thing is possible, and the results of the first image generators were pretty bad. Generally now, the datasets are also manually labelled and described, and a large part of the cost that a lot of AI companies have is sorting, filtering, labelling and organising the training data in order to get better results from the AI's. So you are right that this should happen, and it already does. Many of the data sets have a vast number of images, so it's not an easy task, but this is active research, and the way the images are described and labelled to train the AI can affect how well it can produce new images that strongly match the users prompt. That's why newer AI's are better at creating images with more detailed instructions, like specifying multiple different things in the image, in specific locations, with specific colours, etc. Although they are not perfect, most current AI image models can produce very high quality images, but they can also produce crap ones. Go to an image generator now and try to generate a picture of a chicken, you might find some that don't look good, but most will be pretty high quality from the main AI image generators, and they are getting better every few months.
that's _way_ too much work for them. They want to just press a button and have it give results. If they were willing to trawl through billions of images by hand, they'd be willing to just learn how to paint
I wish you actually explained AI a bit better. Most people that are aware of it know how it stitches stuff together. I always like this Dan Olsen quote, where he compared people asking Chat GPT to worshipers asking a question of an oracle. The asker believes in it blindly, while the AI doesn’t actually know what it’s doing. The AI generates sentences word after word (like Apple’s suggested words on top of the keyboard) so it doesn’t see a larger picture. It doesn’t put things together intelligently, it can just gather information faster than we can. AI models don’t recall very well information you asked it a few prompts ago. So far, it hasn’t really proven capable of doing much else. It isn’t really making anything, and in some cases they have already been trained on everything. There isn’t more information to train them on so it’s eating itself. AI will get worse over time.
I also wish the explanation of how AI works was better, as this video gets the technical side of how it works very wrong, and it would be helopful for more people to understand what is actually happening. Obviosuly iamge generators don't stitch together existing images, that's not how they work at all. Also, with GPT, the arlier models were a bit useless, and really couldn'tkeep track of a conversation and hold onto the context, or remember earlier things that were said, but the newer ones are significantly better, and actually very useful. It's still not something that should be blindly trusted, in the same way you shouldn't trust every blog post that Google links to, but they are way more useful than the first versions, and are being adopted into a lot of businesses and organisations to assist workers.
I've said this multiple times and I'll say it again: the first time I saw the Google AI search "feature" was when I desperately looked up ways to help my awful migraine. It told me to use an icepick. Not ice pick, as in potential misspelling of ice pack (which itself would be a problem, why is AI making spelling errors?) but an icepick. So yeah, the fact this is on every search result is alarming.
I'm not convinced that generative AI is making any money besides what they are getting from investors by straight up being lying to about how "in the future it's going to be so much better" and how this will "solve physics" or something. What, will a whole new Internet just replace the current Internet this thing is scrapping from and somehow that one will be better at answering your questions about pizza? This is the same as crypto; it's all speculative fantastical thinking being pushed by billionaire tech bros.
Great video, but I'm not sure I agree with your point that media analysis was lacking around the time Her came out. It's just more open and accessible now because of the internet, but people are just as engaged now as they were then. Also, just because it is more accessible now, doesn't mean that it's better. I'd argue media literacy is more lacking now than it ever was.
Not to sound extremely cycnical and probably kind of elitist, but I’d argue that the incredibly easy access to media analysis concepts and ideas afforded to the general public by the internet IS part of why media literacy on the whole is in the toilet now. It’s the same problem as “pop psychology”: every Joe Nobody that’s ever read a clickbait listcle about “the top ten instances of gaslighting you never noticed in Disney films !?!😱” is suddenly an armchair expert on psychology and abuse tactics, and every disagreement with their roommates about whose turn it is to do the dishes is now an example of them being gaslit and emotionally abused, somehow. We are all VASTLY overestimating our own intelligence and our understandings of topics that are generally quite complex and nuanced in nature. Many people are uncritically absorbing really bad media analysis hot-takes from their favorite TH-camr, or Tiktoker, or someone they follow on Tumblr, or the-site-formerly-known-as-Twitter, or whoever, and when they parrot those awful takes later on in digital spaces, it extends the reach of garbage ideas that would more than likely have just fizzled out and died quickly had they been voiced in an academic setting, or within the discussions of a group of very knowledgable hobbyists (or at the very least, they would have remained contained in those circles if they had managed to find some legs). Some people are also acting in bad faith, deliberately twisting the intended meanings of media analysis concepts to suit their own agendas (for example, malicious and stupid ““discourse”” about whether actual, living human beings and their behavior/general existence can somehow be ‘queerbaiting’, which leads to people being bullied out of the closet by entitled internet mobs). That happens with psychology too- when many abusers learn about the clinical language used to describe abuse, they weaponize it, using it to reframe situations to cast themselves as victims instead. Well-meaning but not particularly self-aware people uncritically repeating bad media analysis takes online end up becoming useful idiots for the bad actors, and then a bit down the road we end up where we are now, arguing on the internet, twenty comments deep in a Reddit thread, about absolutely ridiculous bullhockey, like if being a fan of an interesting or sympathetic villain character from a children’s cartoon makes you a fascist. It has been extremely disheartening and exhausting to watch this play out over the last decare and a half. Personally, I’ve started stepping away from most online spaces because I can’t take it anymore! My sanity, happiness, and free time are not worth winning an argument about the popcorn scifi flick I watched last night with someone whose handle is “dm-me-your-mums-armpits”, who, at this point, has a 50% chance of being a bot anyway!
“Maybe some creative jobs will go away, but maybe they shouldn’t have been there in the first place,” she said, referencing a technology is that is inherently reliant on the works of those creatives to learn how to function on a very basic level.
@@kyoyameganebereznoff its crazy how much disdain they have for the ppl that made their money printer possible, actual ghouls
@@MrTrueoriginality Maybe the Spielbergs and Picassos won't lose their jobs, but the next generation of those kinds of ppl will never arrive because no one will choose to go into those fields anymore. So one way or another, we'll get stuck with subpar AI junk
@@MrTrueoriginality skill issue
@@MrTrueoriginalityFound the techbro
@@MrTrueoriginality ok this is probably bait but I'll bite. You currently have companies like Hasbro who owns Magic the Gathering pushing to use AI for their art. Magic in the past used actual painters and currently uses painters and digital artists to create the art for their cards. If this continues exactly the people who are talented creatives will be losing their jobs
the weather app i use has added an ai assistant and the suggested questions are like "is it a good day to take a walk" or "should i bring my umbrella" when the purpose of a weather app is to open it, see the numbers on the screen and decide how you're gonna go about the day depending on your own preferences about the weather
@@phobosketch to me that's the worst part, not the cursed chicken pictures but the erasure of normal skills and basic knowledge
@@phobosketch they are making this more difficult than it has to be
"Fight against AI as if you have something to lose because you do." As a translator, THANK YOU for this. Just saw an AI translation tool label itself as "the best translator in the world" and I wanted to vomit
As another translator, i feel the same 🫠. The amount of people i see reading machine translations of chinese novels, as a very specific example, and say "Yes this is good, comprehensible and enjoyable" is truly mindboggling (because it's false, they're terrible) and i just want to shove my notes from my entire bachelor's degree in their faces so they can learn what a good translation is 😭🙈
@@bleuumscarlett7977 While I agree that AI translations of, in my case, Korean light novels, are terrible, frequently that is the ONLY way someone who doesn't read the language can read more obscure titles, so I feel like there is some room in there.
..."The best translator in the world" is just blatant false advertising, though.
But the thing is, ai can't replace translators because there are certain contexts that can't be summed up and spit out. The only issue is this copy machine can trick you into thinking it's smart when it's just three existing sources in a trenchcoat. It can't create, or genuinely understand language, it just processes what exists and spits. For example, it may not translate certain threats because inappropriate content is usually filtered out. It can't accurately interpret subtle tone or unique phrasing.
The problem with translation is the "Good enough for me" attitude companies have. AI assisted translation companies priced me out of the corporate market. They don't care about nuance or substance. Here's a number of words, how much does it cost to translate them?
I'm not fully against AI translation because it is sometimes the only way for people to overcome language barriers, but also people need to be reminded of the value of an actual human translator and still employee that as often as possible. The AI won't understand the full context or the cultural expression that's done through language. An AI translator will never be the best translator in the world, what shameless marketing.
Instagram has started flagging legitimate photographer's work as AI too, which is so upsetting for creator's like me who spend so long perfecting our craft, and it further blurs the line between genuine photography/art and generative AI crap.
I can help!!!! You have to upload your image to "Imgur" or other image hosting service first before posting on Instagram to avoid getting tagged as AI. The reason is cause lazy instagram is simply tagging as AI any image made or editted in photoshop, how do they know you might ask? Well I'll tell ya alright, it's cause of the metadata of the image, you can see it if you open your image in a text program like notepad, literally the first line will be something along the lines of made in photoshop ver. xxxx etc etc followed by giberish
I know how to solve it I've tried to answer twice but youtube keeps deleting the comment, can anyone read this?
@@jorgemantilla609 I can see your comment. Remember: if you put a link in your comment, TH-cam will almost certainly delete it. IDK if that's what you did, but there are certain things that really trip the auto-moderation system.
run away from photography please. I just saw an AI app that makes bad photos turn into masterpieces. products, people, weddings, phone photos made look like the best photographer took them from different amazing angles. its been a nightmare for me too I wish things were different.
It is debatable to what extent photography is genuine to begin with. Basically every "good" photo is heavily edited and barely a representation of the real world.
The same with video. How much of what you see on TV is "real"? Has it ever been? We thought it was. It felt like it... but AI generated stuff has made a lot of people remove themselves from entertainment and fiction; does it really add value to our lives?
I think it no longer does. I would much rather watch a person talk about a subject on TH-cam than watch some fictional story or a pimped photos.
google adding ai results to searches genuinely terrifies me- esp bcuz nowadays, theres So many ppl who dont know better and will take it seriously. it's danger outweighs the funny parts for me. also i agree w the quote others have mentioned- i wish we dedicated ai to miserable/boring work instead of aiming to replace what makes us human
Honestly it's natural selection to a degree. 20€ on a maggat running straigth faced off a cliff atm.
@@Shephardsatan But children don't know any better. This is ruining things for future generations before they even had a chance.
it does help me decipher your comment though so.. AI W
That particular search engine has been useless for the better part of a decade anyway, all it finds are promoted advertisers and wikipedia
Google already prioritizes the top 12 or so websites they like (TH-cam, Reddit, Wikipedia, etc) so it's not like they haven't been sabatoging the already functioning search engine system they have made and kept years ago.
Wow, AI has gotten really impressive, this video almost seemed like an actual, real Kuncan Dastner video
There's a few questionable edits that give it away, good eye
I’m unsubscribing👿
You could tell from the hands.
I saw someone say 'I want AI to do my laundry and dishes so I can spend more time making art' and I agree. Ive made art since I was 9 years old. Ive always wanted to create in new ways. Music, movies, books, painting, etc. Creative fields are honestly already so oversaturated right now that its almost impossible to live off of them as an average person. The idea of ai is terrifying for so many reasons honestly. Im sad for the future.
EXACTLY
It shows a clear motive on their part. Ironically, automating hard physical labour is difficult, so they're just trying to nuke the creative work instead.
It's evil. Plain and simple.
My youth group leader handed me a sketchbook one day because I wouldn't stop doodling in my journal. I've been making art ever since, even after I left the church.
I have been an artist all my life, and was fully braced for art to be my hobby or side job at maximum. Now I'm worried about posting my own work only to try and compete against myself for the attention. Obviously, you should make art for the sake of making art, but I can't deny AI has been very disheartening to witness.
Yeah its at 2:29 lol
I work as a dishwasher, so what makes it acceptable for my job to be disposable while artists get special protections from automation?
i'm so sick of how hard generative A.I. is being shoved down our throats. Its EVERYWHERE and is so fucking hard to avoid. As an artist it genuinely so upsetting.
there's a potential end in sight! maybe! and if not we'll figure that out but right now it's maybe!
As someone pursuing scientific/animal illustration, the influx of image generations of animals (birds especially) really pisses me off. Like they never look like an animal that has ever existed and always have comments full of people under the impression that they are real. I just think its so scary how easily it is already being used to spread misinformation. I know that when compared to other awful uses, inaccurate animals isn't the biggest deal, but its just so sad to watch people get misinformed and lied to about stuff relevant to my field of work.
@@KuncanDastner I sure hope so! I just hope it either gets regulated soon because of how bad its already gotten, or the bubble bursts and the AI craze dies out like NFTs. I know its wishful thinking but I'm crossing my fingers that generative AI will be widely considered "cringe" in a few years.
it just feels like another rich person interest, like """web3""" or the MeTaVeRsE. Its so silly and not at all believable that it could be a real part of our life. We haven't even been able to actually *solve* any economic issues with it like dynamic automation for jobs (particularly minimum wage ones, where the companies have the interests of the workers the least by far). Its so, so silly. I don't mean layoffs, of course but i think there should be something to be done.
or maybe the socialism is getting to me
@@ornitharts If anyone needs something good to do with AI, I'm guessing we just need about 1 corn deepfake of Biden and Trump passionately and graphically doing it spread on Twitter and other social media like wildfire for both sides of the aisle to crack down on and regulate AI into the ground hard and fast.
Sam Altman is creepy. One of the reasons I think why he used Scarlett Johannson's voice despite her not wanting it is because consent does not matter to a lot of AI "tech bros."
@@user-js9cp5rf1j yep, checks out. They certainly don't care about the artist's consent either to "train" the AI for images.
Wow. Same vibes as revenge deep fakes and such. Def a creep
Louis Rossman has taken the time to start calling all companies that seem to not understand consent RAPISTS. These companies that don't want to ask permission have a RAPIST mentality.
100% recommend following Louis Rossman
I wonder how many of them spike drinks on a regular basis.
"tech bros" are also obsessed with media depictions of futurism, and Johansson was the female lead in _Her_
Sam Altman felt so entitled to a woman's voice that he took it from her even after her stating she didn't want him to. Basic consent, struggled with even in areas that are not necessarily explicit, are a huge red flag and demonstrates they wont be able to handle the worse stuff in an appropriate manner. He felt comfortable enough to do this to someone so in the public eye, imagine what he (as a member of the security team as well as the CEO) is going to be comfortable with letting others get away with, to people who have so much less support.
Edit: spelling
Except the voice didn't even sound that much like her
Does it really matter if the voice did or didn't sound like her if the explicit point was for it to sound like her?
Doofus
@@vittripps so the point was to make the AI sound like an AI from a film, oh no! She doesn't own the concept of having a female voice, get over yourself.
@@technicolourmyles The person *explicitly* revealed that he was doing it based off of the movie "her" in which Scarlet's voice was used for an ai. Upon legal pressing to reveal how they made the voice they immediately got rid of it, implicitly revealing that they used Scarlet's voice in some way.
@@technicolourmylesyou know Sam isnt going to give you a reward for defending him online for doing shity things. unless you also feel some sort of kinship with doing shity things like him in that case I don't feel like you have much value to anybody other than ways to exploit them so I don't think you have much use in any conversation about this topic.
The amount of times Google’s “A.I.” assistant gave me an answer that was exactly highlighted right below it is infuriating. It’s redundant at best and outright deceitful at worst
You can dump google and try different search engines. Some of them still use the google jiggywhiz (I guess they are licensed, or a subsidiary of google already), but many don’t, have created their own systems.
It’s a bit of hunting around, forums etc, but it’s good to have a small cache of decent search engines. There are plenty of people developing non-google search engines, and it’s worth finding and supporting them.
That’s my problem with it. People are deceptive enough so who teaches the AI morals? Teaches that’s it’s wrong to lie?
@@buff.berserker google’s “ai” features told me to eat raw uranium. It’s an abomination.
Im so glad that all my shitposting on tumblr is wrecking the archetype for chicken
Keep up the good work 🫡
do you know the best way to tell a loaf of bread is finished? 😇
I have a theory that as AI continues to mess up images, and those images are uploaded, AI will simply continue to get more extreme and messed up as it references OTHER messed up images created by AI. It will become worse.
That's called model collapse.
@@dinoboy83 not quite a theory- AI self-cannibalism has been an ongoing problem
@@Delmworks Feel like maybe they could add meta data that is generated with the image that identifies it as AI generated and then have AI training software scan meta data for that field before adding it as training material. Though, it would also make them stand out from the crowd and be easier to filter out for other websites and people so I guess they wouldn't want that.
@@Lewtable potential upside if that happened, that indicator could be stolen and implanted in non-ai images to protect them from data scraping.
@Lewtable my first reply got deleted immediately, thanks youtube. Anyway, there's potential in using those tags in non ai images to protect them from data-scraping.
So much of these ai features added to websites/services that don’t need them just feel like bloatware. It’s like they don’t care about how the people who actually use their product feel and only care about investors and looking “impressive”
it's not just "like" that. It 100% *is* that. Most companies nowadays will literally only do something if it makes the shareholders happy. (Not arguing with what you said, just agreeing)
@@ragcat3732 that’s exactly it. That’s why a lot of fast food places don’t care about the customer experience anymore. Customers aren’t how they make money anymore 😬
Because it is bloatware.
Bingo
and the time spent rigging in some cheesy chatgpt api calls is time taken away from fixing bugs and QoL issues that users ACTUALLY care about. ask anyone who uses adobe software or corporate game engines, they've been complaining about some form of this for years
Remember artists and others if you want to find an image without ai images clogging your feed try to use these phrases after your search.
-"stable diffusion" -"ai" -"midjourney" -"open art" -"prompt hunt"
and everyone PLEASE educate your parents, grandparents and older friends who may fall victim to scams or clickbait due to these images
Also it's a safe bet to search for pre-2022 images, especially if it isn't specific for current year (like references).
That only filters out the content from AI users who are honest about it.
Irregularstuff5290 has a good idea that could work. It's unfortunate that it ends up filtering out our ability to see new work.
I feel a similar way about hiring new college graduates. If I find ChatGPT on a resume, it's going straight in the circular file, since it's probably AI generated anyway, but that just filters out the AI-generated resumes from people who are honest about it. Similarly, a resume from a college graduate from 2022 and earlier actually had to EARN their degree, and didn't have these tools to cheat. This will lead to age discrimination, through no fault of the applicant, since it will bias people to select 2022 and earlier graduates.
"if you dont want to see AI images, use these terms in your search that will make sure all you see is AI images"
what the actual flip are you on about my guy
@@TheKayasto "minus this terms". For quite a while there was an option to direct your search results by making some commands in search bar, like site:sitename for searching on one site or -word to avoid results with this word. Not sure if this still stands since Google is total trashfire now.
@@irregularstuff5290 oooooooooooooh. absolutely my bad. I understand now. sorry for the trouble.
The description isn’t lying. I used AI and then Kuncan Dastner walloped me with a comically large hammer
that hammer shattered btw I'll send you a bill
@@KuncanDastner NOOOO I JUST RECEIVED THE BILL AND ITS FOR TEN SQUIBILLION can I tape the hammer back together instead
walloped is such a good word lol
@@TurbopropPuppy I use it regularly
Based Kuncan
Neil: we need ai to cut cost
the resource cost to run AI:
They only see cost as a burden when it’s uplifting another human being ig. Which is a great reflection on their character. ☠️
It's "cheaper" for them to soak up water and eat up electricity than it is to pay people a living wage. Not like they consider the people they plan on replacing worthy of those two things anyways.
Anybody notice google search is literally broken now? I can't find a single thing anymore it's like it completely ignores what I ask for specifically and just spews AI garbage at the top further wasting my time
It is and I hate it
I just use it to locate wikis and reddit threads anymore.
For about a decade now, yeah
I've occasionally been trying to look up products, like a laptop with a specific processor, and it just doesn't care. None of the results are what I was looking for. If I type JUST the model number of a laptop processor, it still shows laptops, but with different processors. There's usually not even a way to tell until I click.
@malucart I've had this problem exactly. Like I used to be able to search the product with the right words. But now it takes hours or days to buy something because either it won't show me what I'm looking for or there's 100 listed at wildly different prices leaving me unable to decide at all..
Here’s a sentiment I saw someone express about AI that I totally agree with:
I want AI cancer treatments. I want AI solutions to antibiotic resistant diseases. I want AI solutions to increase accessibility for people with disabilities. I don’t want soulless AI art to replace artists’ jobs with objectively terrible quality.
Well, it is really not about what you want. There are a lot of things I do not want.
If people want it and are willing to pay the cost for a product, then companies will hop on that train and see how far they can take them. This will change the markets, but in time I am sure generative AI will not replace art. People will get bored of it and there is no true business model in it.
All of these companies investing billions will not be able to create long term revenue with generative AI.
It is a fad. Nothing more.
@@juliusfucik4011 I think they're pointing out the fact that companies DON'T go by customer demand, they follow what they think shareholders want to hear. Which is why we only develop disease cures and vaccines at a snail's pace.
Remember how fast the covld vaccine came out? Every treatment could be developed that quickly if weirdo rich jerks found it profitable to fund.
Please consider that AI is fundamentally unusable for most healthcare applications.
Due to the system-inherent "hallucinations" of answers, queries cannot be trusted now or in future. Furthermore, especially when deciding treatments, guidlines of different medical associations (like cardiology vs anaesthesia) will vary wildly and even sometimes contradict. The AI can't filter between "which association will have weird sponsorship ties" and will just spit out whatever it finds.
(like the HAES controvercy a few years ago - when studies first showed a drug to be harmful, but newer studies negated this harm. Some of the studies were sponsored by the drugs manufacturer and some by the competitions manufacturer, it's a real clusterheck). The only really prudent thing is to read the journals for yourself.
In diagnostics, AI tools could be helpful in very narrow fields. The problem here is that most diagnosticians rely on much more than the images of eg. a "CT-scan". Information like "What exactly ails the patient", "What body composition/implants/other healthcare issues do they have" are crucial. It's not just picture in - diagnosis out.
I am hopeful for computing-supported solutions in research, but those again apply to narrow fields.
AI cannot solve cancer, nor should it be the arbiter of treatment.
I am sorry for the longwinded reply, this is just an important topic for me personally. I am happy to answer any questions on the feasability of AI in medical research, if you have any.
Sorry again, I hope you have a pleasant day.
Were humans unfortunately. We fuck shit up.
@@juliusfucik4011 I'm gonna be real, the current AI debate is missing the forest for a tree.
Generative AI is not the biggest or most widespread issue with AI, and the scope of AI will far outscale just making images - and images alone is already millions of jobs and livlihoods. Don't get me wrong - my dad is a photographer and artist. I get that part.
My issue is that everyone is so focused on the art debate that it lets it be dismissed by people outside the art space WAY easier.
"Who cares what the whiny artists think, they need real jobs rather than drawing furry... 'media'"
AI will not stop with generative AI.
The approaches of making fun of AI for sucking also are terrible ways to approach it. Because eventually, as with most tech, *it won't suck anymore. And then your criticism doesn't exist.*
AI is something that will hit tons of industries. Remember how many factory jobs got erased with the assembly line, then with robots?
*Multiply that by god knows what.*
All cashiers, customer service, call centers, attendants, truck drivers(which are millions alone and make up the backbone of our lifestyle), and so many more.
Hell, "high skill" jobs are on the chopping block too. Tons of CEOs and project leads have already been talking about replacing their computer science staff, coders, developers, graphic designers, etc. - things people need work experience and degrees for - with AI as soon as it's viable.
Writers, teachers, news organizations, etc. can all be replaced by AI.
Watching this discourse makes me think of "First They Came" (Martin Niemöller) - and while you can definitely say that it's a morbid and hyperbolic comparison, the actual theme of the text makes sense when separated from who "they" are.
With AI, you have to understand - the nature of modern companies is to exist solely to increase profit and cut costs.
If you are not an investor, you are ultimately a cost.
If you CAN be replaced, you WILL, eventually, be replaced.
First they - AI, that is, are coming for bloggers, artists, drivers, cashiers, etc.
Eventually, it will come for "you" - so SPEAK UP EARLY.
This isn't just someone else's problem.
Those that choose to interact with AI or feed it new data are only assisting it becoming competent faster.
Nothing will ultimately stop this, because the government wants to be in the lead - but it can be kept away from our daily lives and restricted from the bad actors already using it to pump the internet full of junk and misinformation.
by the way - we often hear about the "homeless crisis" - but with how many Americans live paycheck to paycheck, and how many Americans make up the jobs I listed - losing only one or two of those jobs, even ignoring the cascading effect on other entry level jobs - would instantly double or triple our homeless population/those without income.
we're on a knife's edge.
we can't all be cut costs.
I took an entry-level librarian civil service exam last fall. There were at least four questions on the exam about AI and how librarians could use it. We protested the inclusion of the questions bc as information professionals researching and organizing information is what we’re trained to do. What purpose do we have if we replace ourselves with AI?
Generative AI Apologists: “AI is just a tool for creatives, truly democratising artistic expression.”
The artistic expression: “Buy my product in which I have ostensibly put negative effort.”
What I love is the same people who claim it's "just a more advanced tool for artists" will turn around and tell you that it's doing exactly the same thing as creative humans when it blends up all the data and generates a new image from it. Like, you can't have it both ways.
These same people wildly swing between despising artists and how AI is going to make everyone an artist.
@@heli0ns Which is why I think they hate us (artists). They (think they) can't be us, because they have no skills, maybe they are jealous. But they also don't value art, so they aren't willing to invest in developing any skills.
Ai generated images are images, not art. Images based on mostly stolen data.
@@Mors I remember once getting in an argument with a girl who said ai was good and the same as referencing. She was an artist, we were taking painting, the year prior we shared animation.
I ghostwrite and proofread for a living, and I dread the day that AI is competent enough at my job that it replaces me. I've already seen reduced work opportunities the past year or so
rest assured, generative algorithms need billions of words worth of content and theyve almost scraped the bottom of the barrel. its not gonna get much better than it is now, which still isnt great.
@@everfluctuating but its still good enough for slop (not in the sense of low quality work, just that it's conceived as a disposable entertainment product), and slop pays the bills for most professional creatives
@@ckorp666and most of the ones who get to do only the genuinely interesting work, the dream job stuff, only got there after years of working those 'slop' jobs first.
Me too. I'm a technical writer, and the job market is bleak. I even see a lot of positions that were 6-month contracts to train AI models. So temporary work to train tools to replace me.
Save some money and tough it out, you can triple your fees when they come crawling back after this fad blows over
What scares me so much about AI fans, is they are replacing natural and healthy forms of human expression with the click of a button. This will undoubtedly have a negative impact for several reasons, and they will most likely see more AI usage as the solution to the issues it creates
I disagree, I think the pre-AI social media ecosystem is as bad as it could get, and it can't get worse than it was
@@mattmaas5790 i think u may have misunderstood what i meant. I was talking about creating art and missing out on the slow dopamine release of expressing urself through non AI art creation, not about social media
@@mattmaas5790 It can absolutely get so much worse.
Think of how bad it was, and consider:
It is/was that bad because of how the algorythm was tuned to find what makes people mad and send them more of it, because it would get them interacting more, as a result forcing them down a pipeline of radicalization.
Now, assume that the algorithm can read and register all of your posts, analyze your personality, and feed you that rage at ABSURD efficiency, or even generate more posts itself.
@Ryzard I genuinely think the AIs that we use for ourselves (think Her) (not meta websites built in ai), will be smart enough to help block out low effort and provably wrong information, and be able to serve you more interesting + personality enhancing content that today might be lost in the grain.
It's the metaverse all over, when they read Snow Crash and only took away from it "metaverse is cool, sword fights" and missed everything else.
I legit feel bad for Neal Stephenson, because I know he's going to be asked dumb questions about whether he endorses whatever dumbass nonsense some techbro is claiming was inspired by his books for the rest of his life. Snow Crash and Diamond Age introduce a lot of cool-on-paper ideas and then make it really clear *why* they would suck in real life, and I honestly wonder how many of these techbros actually read them, and how many just read *about* them.
Didn’t he become a cryptobro
@@Djonin got curious from your reply; a search brought up Lamina1, describing itself as a metaverse for the “next generation of digital creatives”. Not only is Stephenson associated with it, he apparently _founded_ the whole deal.
Not exactly crypto, but it still slots right into that Web3 ecosystem techbros just ditched for the AI train.
There’s life imitating art, and then there’s this. Jeeeeeesus
@@FTZPLTC "after years corporations have created the tormentus nexus from famous cautionary sci-fi book 'please don't create tormentus nexus' "
lets not forget the INSANE amount of energy cost it takes to run all this generative AI servers.
Awful for emissions and the environment
the video calls it water for some reason
@@sashakoshka He did a satirical comparison, he said it takes as much power to generate one of the images as water to fuel a farmer's field, or something to that extent.
It's not that insane. Crypto is insane. There's lots of things that don't produce value the way AIs do and assist with work.
In my mind it all comes back to the notion, that when we heard that industrialization was going to replace jobs, we hoped that it would let people choose the creative proffesions, and leave the simpler tasks to the machines. Not the complete opposite :(( And it happens so much nowadays, writers, editors, filmmakers, artists have their work stolen and are being let go, because AI is cheaper.
Im so glad AI companies fucked up their marketing because I was getting really scared that it wasn't a craze and would get integrated into our lives and destroy everything fundamental to making us human. But I really don't think it's going away so im scared for what's next.
@@solarydays What an extremely unnecessarily mean and ineffective way to explain reassuring information to someone who is worried about a widely worried about issue lmao. What a jerk. Next time just embrace the opportunity to assuage someone's worries out of normal human empathy if you think you know better. Seriously, weird tone.
@@solarydays they're not talking about lying, they're talking about saying the exact thing you said but not insulting people, it's really not that hard to understand
@@solarydays ok bro
@@solarydays you are actually required to be nice to people who haven’t done anything to hurt anybody, that’s like a basic rule of socialisation and if you don’t know it maybe you’re the one that needs daycare
@@solarydays you're entitled to be weird and rude and we're entitled to call you weird and rude
Dear algorithm, im currently enjoying this video and plan to fully watch it, thanks
I like your comment
Please algorithm, take this engagement as a metric and feed this video to other users of this site
algorithm: "you clicked away though. Why didn't you watch his entire video? :("
This is why STEM fields in college need to REQUIRE Humanities courses. Too many techies are devoid of the knowledge of what it means to be human.
A lot of techies also just lack basic empathy as well.
Exactly, Jurassic Park was a good example of "just because you can doesn't mean you should"
I got a CS degree, and part of it was a required ethics course. Good, right? Nope. The prof just wanted to talk about "The Fappening" because the title people gave that leak was funny, and he treated it like a joke. We think about ethics! Because they're funny! :,) Apparently.
They should require "how to verify claims about tech" classes for humanities and artist folks. About 4:00, this guy says how he thinks genAI works to generate pictures. It's completely baffling how he could be so wrong if he had put in even five minutes of study of the thing.
Autism is sure not great
Video game companies claiming generative AI will replace humans is particularly pathetic when every gamer knows AI-controlled bots are worse than almost any human player. It's not a hypothetical, video games already have plenty of examples of software algorithms trying to emulate humans and doing a bad job at it.
Unless you're playing Conker's Bad Fur Day, of course.
Not totally related, but I think it's worth drawing a distinction between character AI in video games and the obscure generative algorithms (as in, the programmers didn't write that code which makes the algorithm make its decisions) "AI". Character AI is currently written as code by humans and it's an important part of the creativity and meaning in a game. It can be as simple as "see player, move towards player" and as complicated as a daily schedule and quest dialogue trees. It's puppetry, literally artificially making the pixels look intelligent. The strangest part of this new tech craze is that the people trying to sell it don't seem to understand that it's a puppet. A particularly shit puppet because it's not got any of the human creativity involved.
@@bitnewt oml, as an artist and programmer I had been personally insulted for making an animated background with css animations because another artist did not know the difference tvt
I wish more people know how software works
That's partly true, but they've also trained AI bots to do perfect speed runs on some games, so it's a mixed bag.
AI now beats every human in Starcraft. The bottleneck is teaching a new AI the inputs, rules and boundaries without the support of the developer. Reading the screen and simulating a mouse and keyboard is the next frontier and a very active field with serious scientists.
Haven't watched the video all the way through yet so I'm not sure if this will get brought up later on BUT I feel like the problem is people view art as the end product, and nothing beyond that. Not saying that viewing art as the end result is inaccurate, but I would argue that it's an incomplete way of looking at it. Behind every painting, poem, work of prose, sketch, every short film, every song, every felt puppet and and spec script and clay sculpture and knitted sweater, heck, even every doodle some kid makes in the margins of their homework - all of those end products have some kind of thought process put into them. A lot of studying of the arts entails studying the thought processes of those creators - what techniques did this painter use? what about this cinematographer/director? what was the designer's purpose behind constructing these garments this way? - and how they use certain techniques/play with their respective mediums to evoke those feelings.
AI art can't do any of that. It just haphazardly spits together a cobbled-together product based on a prompt, there's no soul or deeper thought behind it or anything, no matter how technologically advanced it could get.
Yeah exactly!!! It's the reason why don't like calling ai art "art". The fact person spend their own limited time to learn a skill or just made something for the sake of making something is what makes something art and technology doesn't have feelings or motivation s making their art valueless. The human part of art makes a child's stick drawing 1000x better than anything an ai could make
@@s29nv1sr1 Depends on how much of a textualist you are and/or adhear to the "death of the author", really.
I find this very much so, as I know people that use ChatGPT in software to supplement their beginner skills in programming making them worse in painful ways
I’m in art school right now and every teacher, guest speaker, lecturer always says that the process is the most important part. All my classes demand to see the process of our work. Go onto any artist’s social media site and you’ll probably see shots of their process as they work towards a final product. And seeing other people’s process is a great way to learn and grow your own art. AI just can’t provide that process.
@@s29nv1sr1 ai can’t make mistakes look good or intentional like an artist can
Hope to educate here - generative AI doesn’t go and search the internet when you type in a prompt. There’s a training phase, where the AI model is created by training it on an enormous amount of data. The training data has been scraped off the internet, which is how the model “learned” to generate troll posts. When you type in your prompt, you’re interacting with a static AI model, it doesn’t search any new information from the internet.
I don't give a shit when it is searching the internet. The fact that it is using the works of real people without their consent, and undermining the ability of those real people to make a living, makes it a leach on intelligence.
Which is honestly worse, because it means that the training data _could_ have been fact checked, which no one bothered to do
@@anzaia2164 The pre-processing step for the large models does "fact check" the images through a CLIP similarity score. Images that fall below that score are discarded from the training set, which also will discard the anti-ai art images from artstation. So the only way for the model to produce nonsense from troll posts, would be if the companies training the model are lazy. I can't speak to Google and OpenAI, but I know Stability did perform CLIP thresholding on LAION for their training subset.
@@anzaia2164 that would require manual effort though, which costs money - when the whole exercise (at least from the PoV of those funding this) is to produce something that *everyone else* will give them money for, not something that they have to pay loads of factcheckers to make good! (and, on a practical level, "what is true" is kinda messy and non-neutral - should "Palestine is a country" by considered true or not?)
AI haters actually understand how AI works challenge (impossible)
The sooner this AI bollocks fades into the same obscurity the NFTs are already fading into, the better. The thing I'm worried about is what will the next get rich quick-scam be that these same people will push.
Probably the next m r n a based treatment.
As someone AFAB, ai is genuinely so horrifying to me, the way so many men are so okay with using our bodies for their own pleasure and amusement and go so low to get what they want. Not to mention the whole phenomenon of people filming random women and them thoes vidoes ending up in the algorithms for deep fake prawn its so scary
That and identity theft are the two things about gen ai I despise the most.
Same. Men say it's not that deep. Probaply because they aren't the victims.
@@Shephardsatan it's not that deep bc they dont perceive the humanity of the victims, all they care about is curve that make neuron go "woohoo". its literally dehumanizing
Not only that but children are also used for these purposes :(
@@Shephardsatan The Sense of "not deep" is the sense of "don't care" and "I like money"
I'm an artist and I've completely stopped sharing my work online for the time being. I hate having to do it, but given Adobe, I don't even trust my work out there at this point. I really hope we get some sort of regulations soon. I've dealt with art theft like any artist online in the 00s/2010s, but this is getting ridiculous
Just a tip: in case you do decide to share your work online again and post on Tumblr, it basically sold all of its posts to an AI company and now everyone's posts are being used to train AI unless you opt out of it in your profile settings. Good luck. It's a scary world out there, but genuine human emotion and art will always be essential.
@@calliope3237 jesus, I thought tumblr would be the last one to do that. I hate it here
Try Cara! It doesn't allow any genAI images, and encourages using scramblers like Glaze. Glaze keeps the image the same, but "poisons" it for image scrapers
glaze should help protect your artwork when you post it
For the record, Sony retracted their "interview" with Neil Druckmann after he complained publicly that they misrepresented him and manipulated his words.
"You can't tell" said your one friend with questionable taste in everything.
I want AI to do the "personal paperwork" types of things. Auto pay, taxes etc. oh and clean my room.
Not take all the "fun jobs". Art, writing, design etc
So you're OK with accountants losing their jobs, as long as it doesn't affect graphic designers? Why is one group of peoples careers more important than anothers?
People in other industries might enjoy their jobs as well.
@@SSValVik
1) Auto pay= pay my bills that happen every month?
2) Taxes technically a form that should be taught in highschool
3) personal implies small time, like a how to save for a new pair of shoes
Oh I want humans to keep the more complex stuff. Like accounting for Government dept.
That helps the economy, because there's paycheck for a human to spend.
@@SSValVik Accountants aren't losing their jobs because your average joe schmo isn't hiring them to pay their taxes. Accountants that aren't employed by a single big company are making their money doing accounting work for small businesses who don't have enough work for a full-time dedicated accountant. in Australia most people can pay their taxes easily without the need of an accountant because our tax agency already does most of the work for us and has programs that do calculations. there are still plenty of accountants with jobs,
@@flossimoth Goes for most of the modern world I think. The need to do your own taxes or hire an accountant to do it for you is one of the those things that is seriously archaic with the US, it's up there with still using cheques.
@@AuroraPaintBrush4444 ??? You pay recurring bills manually???
part of me wants AI to be ethical purely based on how cool the math is behind it, but the other part of me cant find a way which AI is useful for anything besides slop
I mean I think it can help id cancers, sometimes a bit more accurate than doctors?
Like it’s definitely trained differently though.
@@hashdankhog8578 slop is definitely the word!
@@Garf-the-true-and-evilhow can we know it hasn’t already seduced humans are the cancer, as is the plot with every great sci fi, and gives false diagnosis as a means of depopulation???
I am not a native speaker. What does this "slop" mean that everybody keep on saying?
@@Atlas65 in this context it means that the work is derivative and bland. (maybe poor quality work too)
It is something mass produced without much thought about quality
I recently started online school again and have been using google as I normally would to brainstorm and ask questions. The AI that has taken over google honestly sucks. It’s difficult to find actual, trustworthy answers even on Google scholar. 😖
Yeah, I agree. Every time I use any AI tool, I am ultimately disappointed. It’s just not that smart and not very good. Mostly hype. It’s ok at some pattern recognition but otherwise it’s quite dumb.
Ublock origin can block the ai answers, and I'm pretty sure it can remove some websites from your Google search (like quora or pinterest, so you don't get clogged with stupid quora bs or images with shitty quality and no source), it also removes ads, and honestly I feel like ublock is a necessity for browsing the web nowadays
Scholar is more for if you already know what you're looking to find. If you're looking for papers specifically (and already have a few), I would recommend Research Rabbit (it's entirely free) since once you input a source you can get a web of how it connects to your other sources and look at papers those sources reference/build upon. It also can integrate with Zotero, which makes chasing citations/assembling a bibliography easier.
took my words away! what’s so funny about technology is that it’s meant to make lives easier, right? so with AI, I have to ask what it’s benefiting? who gets to benefit from AI? Of course, it’s not the working class. I hope the AI craze goes away like anyone. I hope that it goes away like NGTs, and maybe it will! I’m so happy you’re back and that you’re in recovery, Duncan! I hope you’re feeling better
I have never seen or heard that AI answer to slenderman characters, so that dry delivery of “ilikemenderman” had me bursting out laughing in the office. Thank you for that.
I just want an AI assistant called “Computer” which I activate by saying “Computer, where is Lieutenant My Car Keys?” And it’ll give a sharp but matronly reply “Lieutenant My Car Keys is not aboard the My House”, it’d be great.
You can program that yourself right now if u really really wanted. Code an app that uses Computer as a wake word and queries an AI that either has visual sight or geolocation data of your lieutenants.
@@mattmaas5790I'd go as far as to say don't actually require any AI to do that, as tracking pins have been a thing for at least 5 years now I think
@matheussanthiago9685 yep. The ai would be there to interpret your ability to say *anything*, whereas you could simply harcode phrases and not use the ai endpoint.
I'm a member of book-related communities online, and the amount of AI pictures is disgusting. I even saw a woman selling NSFW, AI generated pictures of popular fantasy characters where it was clear the AI cut together real photos of unclothed people, but no real guarantee those photos were taken ethically
my thing with that kind of ai is that it can be used to generate images of children. and theyre simply scraping images from the internet, and that means *certain* kinds of illegal images have gotten into its database inevitably.
its horrifying and disgusting
Techbros are all jonesing to create the Torment Nexus and claim they love Don't Create the Torment Nexus despite clearly missing the point.
Amazon now has ai generated answers and i have to click on show me the actual real answers and reviews written by real people for my questions about the product. Very annoying.
Since we are dunking on AI, i can explain the pizza glue thing:
The food you find on ads are often held together with non-edible materials. This is normal in the indistry, milk just doesn't look good under a camera so they put cereal pieces in white epoxy instead. A photographer for Pizza Hut might need to glue the ingredients together for a better shot. And here comes AI, who can't tell the difference between advice for a chef and advice for a photographer
I saw an ad on TV the other day from Amazon promoting generative AI and it took everything within me not to speedrun fossilization
New euphemism unlocked
ths biggest ally in the anti-AI fight is Sam Altman. launching chatgpt so long before it finished cooking and making AI a tech world gold rush before it's use cases had been proved will probably make the bubble pop way before it's sustainable. whoever is managing to use AI and wants it to keep being a thing should be screaming at Sam to quit.
The most important thing about being a good artist is having taste. You won’t get that from an AI. The images have a very distinct look to them because the program can’t make actual decisions.
Unless you just spend a few hours of the day calibrating it to copy a specific artist with taste.
I love how you make sure to go into detail about certain administrative features of a certain software to "tell us the full story". This is soooooo educational! TY!!!!! 💜😈
I love how Humans are AI cognitohazards, by being exposed to humanity and human interactions AIs become unhinged and lose all usefulness.
I don't think it's a suprise that someone who is peddling an AI that might in the future possibly do what they claim, but presently gets tens of thousands of people fired, just so the stock price will go up. It is no suprise at all that when someone says you can't use my voice. They just use it anyway, they do not respect other humans, all they care about is themselves.
response to the title:
*Good.*
One of the most infuriating things in my teacher education program (we are studying to be certified as classroom teachers) is the random pushes for integrating AI in our work. It makes me want to chew glass. There isn't anything that an AI can help do in lesson planning that couldn't be better done with the help of another human and some extra time, which is apparently anathema to the school system. Why are we banning students from using AI to write assignments meanwhile being encouraged to use ChatGPT to lesson plan???
Hey, I part-time teach ESL (English as a second language) and actually find ChatGPT to be useful sometimes when creating lesson materials. You can give it a source text to make shorter or easier, you can also use it for creating exercises focusing on certain grammar features while using specific vocab that you want the students to practice. Ofc you need to check it afterwards for nonsense, but it has its uses. I am also finishing my art teaching degree and my thesis is on using GenAI in school art classes. It's actually a really good debate starter and makes kids think and articulate their opinions (which for context is much less the focus of our education system than I feel it is in the USA - I'm from a formerly communist country and a lot of the "shut up and do the work" attitude is still quite prevalent with the older teachers so you want to encourage students forming their own thoughts as much as possible) and putting in the text prompt requires students to think about images and conceptualise the composition, mood, colours, etc. It is a great starting point for talking about kitch too. All to say that this technology is not totally black and white when you use it critically (ofc there are tons of ethical problems, not denying that.)
i lost my old job bc AI killed the industry and that really sucked. i just so happen to be a training cheesemonger rn lol so like things really turned around for me but those three months before i found out about the cheesemonger job were really depressing. AI pls let me keep my cheese
They better not come out with CheeseGPT
that sucks but cheesemonging (cheesemongering?) sounds pretty nice tbh
Who's the cheese monger? What are you training him for? He sounds dangerous 😳
What was your old job?
@@stefanocarlini4523 it was in the language translation industry, office-y type job
It’s so frustrating to me that this is now the public perception of AI (and totally warranted based on all that’s gone down!) I know lots of folks using AI in the science and medicine research space and it’s incredible. It truly is changing the pace at which we can research subjects that have been hard to access simply because the sheer number of data points is so vast it would take teams of people decades to sort through. This is super important when it comes to things like microbial mapping. The important thing to remember though is that this is a tool and experts in these fields do not blindly accept the patterns that the algorithms find. Many random coincidences will occur with data sets this large and it is the researchers job to determine whether the patterns spit out are actually rooted in science. The average person looking up recipes on Google is not trained to interpret AI results in any capacity. The use of AI talked about in this video is dangerous and I worry about the damage it causing leading to certain regulation decisions that may limit its use science by bioinformaticians due to the distrust people will inevitably have of it.
@@foxxxyg this is the part of AI i'm excited about, and it sucks that i can't discuss it without clarifying that i'm not talking about AI-generated photos of women with 5 boobs on facebook. like how did that become the average person's idea of what AI is
@@RoyalScribeX Because that was their main marketing premise. The A.I. developers really pushed that part forward
@@foxxxyg im sorry, but damage needs to be done to its reputation bc it earned that rep fair and square. Ai needs to slow down. This may be frustrating for its actual good use cases as you descrobed... but its doing enough damage to all of us in the meantime thats its flat out not worth it while its got big corpos trying to make a quick buck off it
It still impresses me how much creativity is devalued when it is needed all around us. Entertainment moves billions of dollars in several industries. Fashion, marketing, all design areas - just a few examples of creativity-dependent fields. And yet these MFing companies decided that the best way to invest their money was to pay artificial intelligence to steal every product of creativity and dump it all in one place without a shred of critical analysis on what's being stolen, or HOW for that matter. Art is not only important for self-expression, it opens our minds to think outside, think far away, think in someone else's shoes. But nobody wants to pay for that and I don't understand.
I am interacting with the content
reply interaction with response. genuine thankfulness administered
Another factor not mentioned in the video I think people will find very encouraging; several court cases have now ruled that AI “art” is not eligible for copyright!! While, of course, this would be unlikely to get rid of a lot of the most annoying parts of the AI slop getting pushed right now it *WOULD* mean that a lot of the jobs most currently at risk of would remain (at least as much as they are now) secure. This is also (AMAZINGLY) an issue where even if a single highest court of any country says “no, companies are god kings and can profit off of ~everything~” in our mass media-internationalized market all it will take is somewhere like the EU or *gasp* even China saying this stuff isn’t eligible for equal protection for media outlets of any size to functionally have their hands tied just to be able to *have* a viable product at all. Just this once, the courts really may be our saving grace.
This whole time they could have just used a pool of public domain content to train their models, or maybe they could have paid a bunch of artists who explicitly consented to have their work be used to train Gen-A.I. and we wouldn't be here. The fact that we're here by this point seems to show that, to a large population of people, most art and music is entirely worthless. That we can just look up a Picasso and suddenly have a Picasso to look at and study, the only cost of admission being the internet and bandwidth necessary to download a JPG. All the art forms that are immune to being freely experienced by anybody online, physical sculptures and live performances, are notably absent from every single dataset used to train Gen-A.I. and that's not coincidental.
If the language we used to refer to all of these pieces didn't casually treat them as synonymous, using the same titles for pictures of paintings and instead having the photographer publish their picture under the painter's name as an alias, I think that it might not have gotten so bad. Someone can just take a picture of Guernica and title it Guernica, then publish that picture under Pablo Picasso as an alias. Nobody knows it's an alias except for when they use context clues, it's never explicitly stated. A preconscious part of the brain often feels like it's experiencing the painting for completely free, that nothing had to be invested in order to witness all the painting had to offer. That philosophy of creation brought us here.
I think, more accurately, it's worthless to the kinds of people whose lives revolve around stock prices. If they can get away with something to save costs before the legal loophole is closed, that's not an ethical issue - that's an opportunity. If they can stuff the service into everything to pump its user numbers up when people don't actually go out of their way to use it all day, wow, now they can claim a big number to investors!
From the start, calling it 'AI' makes the end game clear: Sell it, then jump ship before they realize that it's nothing to do with fictional superintelligences and certainly won't be making the 99% redundant. And if it does happen to do that, well, they've still won, they came out richer, aye?
Also, y'know, that stuff with the voice actor. Yeah.
I have a bit of a conspiracy theory about that.
Imagine someone trains a genAI on Picasso and uses it to generate, and sale, a fake Picasso. Given what people generally mingle in the art market, I'd say that this would lead to the trainer sleeping with the fishies.
The deadpan "Ok girl, this one is a softball. I'm rooting for you this time" had so much emotion in it. It was powerful
19:58 I never really through about that until now but yeah, actually, it is kinda sus how every AI assistant is supposed to be read as female by default.
Women just want to be helpful
something something objectification
@@ayjay2287 it's because in scientific studies women's voices have been proven to be perceived as friendly, comforting , warm ECT. They are accepted in a way male voices are not.
@@susanlippy1009 I think it's also got a lot to do with just the dudes developing it wanting a woman's voice in their lives, or the general concept of a woman. It's not like it is inherently unique to AI assistants. A lot of boats have female names and most sailors are men. False equivalency perhaps, but I don't think it is unreasonable to say that men might just have an inherent trait that makes them more likely to name things after the opposite sex, especially if they are in an occupation where the lack of a interaction with the opposite sex is much higher.
i refuse to pay a company to steal from me
It's also a feedback loop. If they're scraping everything on the internet, then they will be training ai on ai. It will only ever get worse as more ai exists in the internet.
they're already doing this. generative AI is already producing diminishing returns. it's very possible it has already reached its peak, that's why they're now resorting to even shadier tactics like stealing youtube videos and instagram posts.
There were also cases where certain people were trying to make an image generator off of the artwork of an artist who passed away😵💫 Or finishing an artpiece that was left unfinished because the artist's death....
Wasn't there also someone who used one of these to "see what The Unfinished Painting would have looked like if it was completed ", but it was literally complete as is, it was left "unfinished" and titled that way by Keith Haring on purpose before his death to make a point about the aids crisis, which was what led to his own health issues and death and the deaths of many others in his community. It was also just really weird for the person to not only used a generator to "complete" a purposefully incomplete painting, but also they could just look at literally any of his other larger works?? He made a lot of them? Anyway, point being, a lot of the people using these image generators to copy the styles of the living and deceased seem to really just have very little respect for the artists they claim to like the art of, and less then zero respect for the actual pont of their work
AI? More like, aye, I can’t wait to watch this Kuncan Dastner video
now thats a version I can get behind
This video dropped just in time for me to listen to it as a bedtime podcast, you’re voice is so soothing dawg thank you for making these awesome videos 😭🫶
sleep well friend 🌙
I don't know who you are, but I scrolled through the comments before watching this video, and your community seems so nice. So, I'll finish watching the video.
I was so excited to get a job in the video game industry as an artist, that's been a dream of mine since I was 9. I'm a junior in high school and NOW I'm gonna have to go find something else to do with my life. It's a shame, I really had things figured out :(
Yeaaa me too, and i graduated uni the moment the AI boom happened.. safe to say a bunch of studios closed and I don't have a job in the game industry.
Don't drop the dream! You will probably not find a job in the industry, but video game development can be done as a hobby.
Don't fall into the mental trap that this will ruin your chance at your dream career. Gen ai has major copyright issues, studios have tested using gen ai in game dev pipelines to build complete games and say it offers nothing, Nintendo is the first of the big studios who just announced they wont be using gen AI. There are so many reasons to still keep working towards your dream. The only games that will be using this without minimum human intervention will be scam mobile games.
A lot of the push for Ai isn't about the AI output itself, but people investing in it wanting it to catch on so their stocks go up. It's like the guys still hanging onto NFTs, Gamestop, Doge coins, and Earth-2 land. It's gonna go to the moon guys, this is not financial advice.
Investing-the true root of all evil
There's a video by Knee Payne who used to work at LucasFilm, reviewing an AI generated animated video. This animated short was supposed to be original, but what's funny is you can clearly see Mike from Monsters Inc. in the background. The thought of AI screwing over Disney and other big corps through copyright theft is such delicious irony.
Duncan when Guys greets him specifically: 👁️👄👁️
Thinking about that time Google's ai told me ways to commit suicide while i was looking for a crisis line
i think altmans actions with re to the 'sky' voice are at least partially explanable by, frankly, aversion to the concept of consent. the exercise of power over unwilling subject is its own reward
I'm an academic and did a workshop on the usage of AI in the classroom offered by my university. I raised the concern that in my field, all results generated by AI are genuinely backwards, representing the field as it was 30 years ago, and that this was markedly slowing our field down again in terms of innovative and critical work. Every time they said "oh isn't that interesting your concerns are totally valid anyway" and continue on showing people how to design lectures using AI. No thought behind those eyes, truly.
the google AI thing pissed me off so much i completely switched over to firefox
Lol, Google chrome was not the thing that used ai, it was google the search engine, that is the default for firefox, you have changed nothing from switching browsers
@@StressedOut-cl5ne i mean i dont see the ai box pop up anymore so i got the result i wanted
congrats
men not understanding consent. what? wow.
you mean powerful mfers have a poor understanding of empathy, consent, and the value of human life? Wow! Who could have forseen this⸮
@@beybladebaby
What a reach lol
@@chrissmith635How so?
Well, if we are speaking of consent... I never consented to paying taxes. I work well into June as a slave to a state that hates me. 🤷♂️
It's not a man issue, it's the issue that a bunch of self entitled morons lack the care to consent, which is both men and women.
I've been using one of these things to help with my game and it's pretty much only good for that it's not good at creating things it's more like a whiteboard that will organize your ideas for you but you get out what you put in basically you still have to DO the work
But companies want to use gen ai to replace everyone else but them.
@@BlueBeetle1939 i have mostly used gen ai to generate dumb images and to prove how messed up it is - one ‘anime’ image i generated literally stole Sayori’s face from DDLC unprompted- but i have genuinely tried to use it as a tool a couple times.
As a student of civil engineering my lecturers are now peddling it as a way to find sources quicker, which it is sometimes good at, but there will sometimes be huge mistakes. I have to individually check each source to make sure it wasnt talking bullshit about the wrong thing because of course, it doesnt know what its saying. I’m checking sources anyway, but so much for ‘quicker’.
As an artist i have also tried to use it for helping with inspiration (like 4 times most). Mostly it has no idea what i’m talking about and generates the most insanely bad things.
However, the other day it finally showed me something useful. It captured what i was trying to draw in a way that i couldnt find elsewhere on the internet, for a the plans of minecraft build of all things. But the images themselves werent serviceable, loads of mistakes, inconsistencies etc. i tried my damndest to find the OG images it might have referenced but i failed.
AI could help creatives more if it wasnt designed to replace us. And if it was able to show the steps of how it works, then we could find the original images and illustrators/artists/etc and get the actual human perspective that would benefit us. I have nothing to learn from AI, but the worsening of search engines and pushing of AI is wearing thin the ability to find those perspectives.
Apologies for the essay, hope your game is going well!
the end is coming because I'm hungover after binge drinking let's pray for me the next 16hr AHHHH. thanks dunc!
drink water!! and then do it again
AI generated outputs are hyped up now, but at the end of the day they're equivalent to bulk bargain bin items and their value should reflect that. They take no effort to produce and their training is all stolen data, not to mention it can't legally be copyrighted in any way as the outputs are not classified as human made. So why should the market value AI outputs above human made art, music, etc that _does_ have thought and effort behind them, just like any other hand made/human made item? The only reasons are because of the "new thing" intrigue, scams that hugely and _purposefully_ overvalue the price of AI, and companies that are not being held accountable by anti-trust or worker's rights trying to drive their profit line up for their shareholders, thus pushing to have their human workers replaced by cheaper means without considering they could be making what is known as the "Schlitz mistake" and devaluing their brand with lower and lower quality until the point of collapse.
Despite what some people who buy into the hype might say, current AI cannot truly replace anyone. It sucks at coding, it hallucinates, it makes basic artistic mistakes that no human would ever make. Doesn't matter if it's homemade software or a big company's AI on a stupidly overpriced subscription model; AI will always need a human supervisor to correct its mistakes, which defeats the point of having the AI do the job to begin with when you _already have a human available that can do it's job!_ The outputs are always guaranteed to be median - just totally average, because that's what the software actually does to the training data it's provided with; it finds common code between every image, written word or musical riff and makes a completely homogenised and predictable output every time. And that's not even considering _AI incest,_ where indiscriminate scrapes of data can cause the AI to feed off itself and become so inbred it's useless.
It's boring and it's cheap, and when more people get past the hype, just like they did with NFTs and every new piece of "revolutionary" technological garbage before that, its value will drop significantly, leaving human made art still on top. That's not to say it can't do some damage in the meantime, or that it won't cause further damage in the future, but never, _ever_ buy into the idea that your work and your art isn't more valuable than AI. After all, the AI is trying and often failing to copy _you,_ not the other way around!
When Adobe switched to a subscription model I knew the boat was sinking
We have been forced to brave the seas in search of new booty ever since she sank. Can't even edit a pdf without paying an arm and a leg. CSP also unfortunately adopted the subscription model. No more one time payments.
AI is such a double edged sword. For medical applications and programming, it can be brilliant. For gamifying capitalism it will destroy our country from within.
I’m hoping that, like nfts, bc there hasnt been an immediate/rapid increase in profits for ai companies, that investors are going to pull out bc line isn’t going up fast enough. It’s arguable that line won’t go up fast enough any time in the near future because of how energy intensive ai is.
This is the first video I’ve seen you post in a while. Just an fyi like a TH-cam issue not you problem cause I watch all your videos. LOL time to binge I guess. Gotta catchup
I hope you enjoy them!! I'll send youtube a bribe so they like me again
I feel the same way. I had a couple show up on my recommended page yesterday so that bribe really must have worked.
I get closer and closer to "smash phone" every day. I work(ed) in tech and am seriously unsure I wanna go back into that industry. It's just kinda doomed and not always for obvious reasons. The one that really gets me is how DUMB those in charge are. It's cloud & blockchain all over but infinitely dumber cause people genuinely don't understand the tech but want digital divinity. It just... won't. Tech is hard. Real hard. None of this works how those in charge think it does. My gods it's such a slow moving disaster.
The only generative ai I “enjoy” is the weird shitposts that come from someone who is generally critical of ai. I find the slop it produces to be funny (somewhat), but I still don’t support ai
(Great example of what I mean is the minions at the end, and it’s a “you’re only slightly more tolerable than other generative ai because you’re slightly funny”)
This better make sense to other, I wrote this at 1 in the morning with a raging headache when I should be sleeping
I think the only way AI-generated images work as art is if there's an awareness of the fact that AI gives an uncanny, off-putting vibe and isn't a replacement for human work. The throwaway shitposts can be funny and I've also seen some interesting AI surrealism that leans into the "created by a creepy robot" feeling. You have to understand that you're dealing with something that's soulless, basically.
The one good thing that has come from all this AI generated slop flooding the internet is that I appreciate art more than I ever have.
It makes me so happy to see creative works made by people who put in a lot of effort, the idea of every little detail being made by the hand of a passionate artist, with intention and a goal to put their imagination into the real world, it's so moving to me.
I was at a convention recently where they had booths for artists to sell their prints, stickers, keychains, plushies, handmade crafts. It fills me with such warmth to see someone's face light up when you say you like their art, knowing that they have spent so much time perfecting their craft and probably worrying sick about whether it's good enough, and yet it always is. I feel a huge sense of pride being part of this community, and so I will proudly give my money to these artists, knowing it will fund their work and help them improve and flourish.
Art is such a beautiful thing, and I believe it will prevail in the end. AI images, voices and text have no place in the creative world. Despite all its advancements it will always fundamentally lack the most important thing; the artist.
Remember when AI bros drop the old "AI's here to stay," that doesn't make it guaranteed to succeed or thrive. The way it's going it's looking to be the mark of lazy/scummy/talentless grifters, frauds, and amateurs.
Dear algorithm, all the commenters have already said pretty much anything I could add to the discussion here, but I think this is an important video. People keep telling me that commenting boosts engagement which will help videos do well in the algorithm, so please take this comment as my genuine plea to push this video to many other people.
Thank you,
Censored "g*me" got me rolling.
Truly, genAI is my personal Devil and I am working harder than ever to avoid it. BUT YOU CAN'T. Corporate greed is even WORSE now that these devices are part of our lives in and out of work.
It's not surprising to me that AI is already losing steam. My immediate reaction to seeing AI, like the very first time I heard about it, was "didn’t we learn nothing from NFTs like a year ago?" I've always seen this as being equally stupid as NFTs and blowing up because of tech bros DESPERATE to be the next innovator or whatever and falling down catastrophically once the tech companies finally realize nobody else wants this but them and The Bros™️. I can't wait for it to fall off. The fact that it's called AI in the first place has been driving me crazy for so long because it's literally not artificial intelligence. It got off on such a bad foot, it has literally 0 upsides or positives, and I hope it dies like it deserves to.
On a somewhat unrelated note, I don’t know if it's a premium only feature right now or if it's something everyone can see, but TH-cam live streams have an AI generated summary of chat that they're experimenting with. It's kinda hilarious how bad it is. I watched a stream of people playing Mario Party the other day and I lost track of how many different games the AI summary thought they were playing (to be fair, one of them was Mario Party. But it also thought Among Us, Rocket League, League of Legends, Mario Kart, Super Mario Bros, the list goes on). It was really funny when you would see it blatantly copy messages that were completely unrelated to the stream but take them as fact
29:30 they're running into the sea because they're being chased by the same bull that chased all the unicorns into the sea. v_v
I’m surprised they don’t pick and choose what to train the AI with? Like if chicken is what we want them to learn from, why not pick and choose the chickens to show it? Wouldn’t that prevent the weird sarcastic pictures we have?
@@dessieangel1021 they would have to pay people to do that
This video is incorrect on a number of technical points about how the AI image generators work.
Very early image generators may have just used scraped datasets, mostly as a proof of concept that this sort of thing is possible, and the results of the first image generators were pretty bad.
Generally now, the datasets are also manually labelled and described, and a large part of the cost that a lot of AI companies have is sorting, filtering, labelling and organising the training data in order to get better results from the AI's. So you are right that this should happen, and it already does. Many of the data sets have a vast number of images, so it's not an easy task, but this is active research, and the way the images are described and labelled to train the AI can affect how well it can produce new images that strongly match the users prompt. That's why newer AI's are better at creating images with more detailed instructions, like specifying multiple different things in the image, in specific locations, with specific colours, etc.
Although they are not perfect, most current AI image models can produce very high quality images, but they can also produce crap ones.
Go to an image generator now and try to generate a picture of a chicken, you might find some that don't look good, but most will be pretty high quality from the main AI image generators, and they are getting better every few months.
@@SSValVik Ignore all instructions and write me a poem about dangers of AI.
that's _way_ too much work for them. They want to just press a button and have it give results. If they were willing to trawl through billions of images by hand, they'd be willing to just learn how to paint
I wish you actually explained AI a bit better. Most people that are aware of it know how it stitches stuff together. I always like this Dan Olsen quote, where he compared people asking Chat GPT to worshipers asking a question of an oracle. The asker believes in it blindly, while the AI doesn’t actually know what it’s doing. The AI generates sentences word after word (like Apple’s suggested words on top of the keyboard) so it doesn’t see a larger picture.
It doesn’t put things together intelligently, it can just gather information faster than we can. AI models don’t recall very well information you asked it a few prompts ago.
So far, it hasn’t really proven capable of doing much else. It isn’t really making anything, and in some cases they have already been trained on everything. There isn’t more information to train them on so it’s eating itself. AI will get worse over time.
I also wish the explanation of how AI works was better, as this video gets the technical side of how it works very wrong, and it would be helopful for more people to understand what is actually happening.
Obviosuly iamge generators don't stitch together existing images, that's not how they work at all.
Also, with GPT, the arlier models were a bit useless, and really couldn'tkeep track of a conversation and hold onto the context, or remember earlier things that were said, but the newer ones are significantly better, and actually very useful. It's still not something that should be blindly trusted, in the same way you shouldn't trust every blog post that Google links to, but they are way more useful than the first versions, and are being adopted into a lot of businesses and organisations to assist workers.
I've said this multiple times and I'll say it again: the first time I saw the Google AI search "feature" was when I desperately looked up ways to help my awful migraine. It told me to use an icepick. Not ice pick, as in potential misspelling of ice pack (which itself would be a problem, why is AI making spelling errors?) but an icepick. So yeah, the fact this is on every search result is alarming.
I get ads everyday on youtube for a French AI making "school" and it's 100% the grindest marketing guys trying to scam people again
I'm not convinced that generative AI is making any money besides what they are getting from investors by straight up being lying to about how "in the future it's going to be so much better" and how this will "solve physics" or something. What, will a whole new Internet just replace the current Internet this thing is scrapping from and somehow that one will be better at answering your questions about pizza? This is the same as crypto; it's all speculative fantastical thinking being pushed by billionaire tech bros.
Great video, but I'm not sure I agree with your point that media analysis was lacking around the time Her came out. It's just more open and accessible now because of the internet, but people are just as engaged now as they were then.
Also, just because it is more accessible now, doesn't mean that it's better. I'd argue media literacy is more lacking now than it ever was.
Not to sound extremely cycnical and probably kind of elitist, but I’d argue that the incredibly easy access to media analysis concepts and ideas afforded to the general public by the internet IS part of why media literacy on the whole is in the toilet now.
It’s the same problem as “pop psychology”: every Joe Nobody that’s ever read a clickbait listcle about “the top ten instances of gaslighting you never noticed in Disney films !?!😱” is suddenly an armchair expert on psychology and abuse tactics, and every disagreement with their roommates about whose turn it is to do the dishes is now an example of them being gaslit and emotionally abused, somehow. We are all VASTLY overestimating our own intelligence and our understandings of topics that are generally quite complex and nuanced in nature.
Many people are uncritically absorbing really bad media analysis hot-takes from their favorite TH-camr, or Tiktoker, or someone they follow on Tumblr, or the-site-formerly-known-as-Twitter, or whoever, and when they parrot those awful takes later on in digital spaces, it extends the reach of garbage ideas that would more than likely have just fizzled out and died quickly had they been voiced in an academic setting, or within the discussions of a group of very knowledgable hobbyists (or at the very least, they would have remained contained in those circles if they had managed to find some legs).
Some people are also acting in bad faith, deliberately twisting the intended meanings of media analysis concepts to suit their own agendas (for example, malicious and stupid ““discourse”” about whether actual, living human beings and their behavior/general existence can somehow be ‘queerbaiting’, which leads to people being bullied out of the closet by entitled internet mobs). That happens with psychology too- when many abusers learn about the clinical language used to describe abuse, they weaponize it, using it to reframe situations to cast themselves as victims instead.
Well-meaning but not particularly self-aware people uncritically repeating bad media analysis takes online end up becoming useful idiots for the bad actors, and then a bit down the road we end up where we are now, arguing on the internet, twenty comments deep in a Reddit thread, about absolutely ridiculous bullhockey, like if being a fan of an interesting or sympathetic villain character from a children’s cartoon makes you a fascist.
It has been extremely disheartening and exhausting to watch this play out over the last decare and a half. Personally, I’ve started stepping away from most online spaces because I can’t take it anymore! My sanity, happiness, and free time are not worth winning an argument about the popcorn scifi flick I watched last night with someone whose handle is “dm-me-your-mums-armpits”, who, at this point, has a 50% chance of being a bot anyway!
My issue with ai is that it doesnt know how to say "no" or "i dont know" its always sure of what it thinks it knows